Rev Chem Eng 2016; 32(2): 149–180

Elena Niculina Dragoi and Silvia Curteanu*

The use of differential evolution algorithm for

solving chemical engineering problems
DOI 10.1515/revce-2015-0042
Received July 10, 2015; accepted December 11, 2015; previously
­published online February 8, 2016
Abstract: Differential evolution (DE), belonging to the
evolutionary algorithm class, is a simple and power-
ful optimizer with great potential for solving different
types of synthetic and real-life problems. Optimization
is an important aspect in the chemical engineering area,
especially when striving to obtain the best results with
a minimum of consumed resources and a minimum of
additional by-products. From the optimization point of
view, DE seems to be an attractive approach for many
researchers who are trying to improve existing systems or
to design new ones. In this context, here, a review of the
most important approaches applying different versions
of DE (simple, modified, or hybridized) for solving spe-
cific chemical engineering problems is realized. Based on
the idea that optimization can be performed at different
levels, two distinct cases were considered – process and
model optimization. In both cases, there are a multitude
of problems solved, from different points of view and with
various parameters, this large area of successful applica-
tions indicating the flexibility and ­performance of DE.
Keywords: chemical engineering processes; differential
evolution; modeling; neural networks; optimization.
1 Introduction
Optimization is an important aspect, the efficient explo-
ration of systems workflow and reduction of consumed
resources being always desirable. As technology evolves,
highly sophisticated and well-controlled chemical plants
*Corresponding author: Silvia Curteanu, Faculty of Chemical
Engineering and Environmental Protection, Department of Chemical
Engineering/Applied Informatics, “Gheorghe Asachi” Technical
University Iasi, Bd. Prof.dr.doc. Dimitrie Mangeron, No. 71A,
700050, Iasi, Romania, e-mail: silvia_curteanu@yahoo.com
Elena Niculina Dragoi: Faculty of Chemical Engineering and
Environmental Protection, Department of Chemical Engineering/
Applied Informatics, “Gheorghe Asachi” Technical University Iasi,
Bd. Prof.dr.doc. Dimitrie Mangeron, No. 71A, 700050, Iasi, Romania.
http://orcid.org/0000-0001-5006-000X
become more available, and the industry is on the lookout
for increased productivity and minimization of by-product
formation (Gujarathi and Babu 2009). In addition, there
are many problems that involve a simultaneous optimiza-
tion of many competing specifications and constraints,
which are difficult to solve without powerful optimization
techniques (Tan et al. 2008). In this context, researchers
have focused on finding improved approaches for opti-
mizing different aspects of real-world problems, includ-
ing the ones encountered in the chemical engineering
area. Optimizing real-world problems supposes two steps:
(i)  constructing the model and (ii) solving the optimiza-
tion model (Levy et al. 2009).
The model (which is usually represented by a math-
ematical description of the system under consideration)
has three main components: (i) variables, (ii) constraints,
and (iii) objective function (Levy et  al. 2009). The scope
of optimization is to determine the values of the system
parameters (variables) for which the overall performance
is the best (measured using an objective function or a
fitness function) under some given conditions (Das and
Suganthan 2011). The variables of the model correspond
to the specific characteristics of the studied system, rep-
resenting components that can be changed to create
different possibilities. The objective function and the
constraints values can be obtained from (i) analytically
known formulae (phenomenological models), (ii) the
outcome of a computational process (such as modeling
approaches) and (iii) measurements of the actual process
(Snyman 2005).
Concerning the problem of solving the optimization
model, different approaches can be encountered. The
traditional optimization algorithms are based on gradi-
ent methods. These methods have a high probability of
getting trapped in local optima (especially in high nonlin-
ear cases) and do not guarantee identification of optimum
(Angira and Babu 2006). As nonlinearity is an important
aspect (as it is introduced by equipment design relations,
equilibrium relations, or by combined heat and mass bal-
ances, among other things) (Angira and Babu 2006), new
techniques that can deal well with these problems have
been proposed and tested (Bastos-Filho et al. 2011, Precup
et al. 2012, Mavrovouniotis and Yang 2013, Yacoub et  al.
2014). In recent years, population-based optimization

Brought to you by | Universita degli Studi di Padova

Authenticated
Download Date | 3/28/16 10:48 AM
150      E.N. Dragoi and S. Curteanu: Differential evolution in chemical engineering
procedures (with biological inspiration) have been dis-
tinguished as powerful approaches that can be efficiently
applied for solving different optimization problems from
various areas. This group includes algorithms like genetic
algorithm (GA), differential evolution (DE), ant colony opti-
mization, or particle swarm optimization (PSO). Among
these algorithms, DE is relatively new, being proposed in
1995 (Storn and Price 1995). It is a powerful approach, in
many theoretical and practical situations outperforming
its counterparts. It is a simple, new-generation evolution-
ary algorithm (EA), developed for solving optimization
problems over continuous domains (Abbass 2002, Hu and
Yan 2009a). Due to its characteristics – simplicity, few
control parameters, and ability to handle nondifferenti-
able, nonlinear, and multimodal objective functions – it
gradually became more popular, in the field of chemical
engineering being used for a large area of applications
(Pant et al. 2009b).
In this article, a review of the most relevant works
applying classic or modified DE versions, in stand-alone
applications or in combination with other specific soft-
ware or algorithms, is performed. The applications dis-
cussed in the next sections are dated after 2009, but DE
has been used to solve chemical engineering problems
from the early days. A few examples are given in Table 1.
The paper is organized as follows: Section 2 presents
the main characteristics of the DE algorithm, its steps, the
stop criteria usually employed, and the principal mecha-
nisms for improving its performance. In Section 3, specific
aspects related to the application of DE for the problem
of process optimization are discussed. A clear distinction
between the use of simple variants (also known as classic
approaches) and the hybrid versions is made. Section 4
tackles the problem of model optimization, and Section
Table 1: A few examples of DE applications in chemical engineering prior to 2009.
Domain   References  
Reactors and fermentation   Chiou and Wang (2001)  
  Wang et al. (2001)  
  Chiou and Wang (1999)  
Wang and Cheng (1999)
Kapadi and Gudi (2004)
Thermal engineering   Babu et al. (2004)  
  Babu et al. (2005)  
  Babu and Sastry (1999)  
  Babu and Munawar (2001)  
  Babu and Angira (2006)  
Fuel engineering   Chen et al. (2002)  
  Babu and Chaurasia (2003)  
Angira and Babu (2006)
  Babu and Angira (2006)  
Brought to you by | Universita degli Studi di Padova
Authenticated
Download Date | 3/28/16 10:48 AM
5 refers to some practical aspects, being a short over-
view that connects the DE algorithm (taking into account
characteristics and possibilities of improvement) with
its application for model and process optimization. The
last section concludes the paper. The chemical engineer-
ing applications are selected to emphasize all the above
aspects.
2 Differential evolution
The DE inspiration source is represented by the D ­ arwinian
principle of evolution, in which only the fittest individuals
survive to the next generations. It was proposed in 1995,
and over the years, as its strong and weak points were
identified, it underwent a series of improvements and
hybridizations. However, the base elements remained the
same.
As with every EA, DE starts with an initial pool of
potential solutions that are evolved (using principles like
mutation, crossover, and selection) until a stop criterion is
reached. The mutation and selection steps (which are sto-
chastic processes) are the main mechanism used for new
individual generation, while selection is (as its name sug-
gests) the step in which a competition for survival takes
place (Greenwood and Tyrrell 2006).
These DE operations have similar terminology with
the ones from other EAs, but their application, role, and
motivation are different. Another difference between DE
and other EAs (such as GA) consists in the parameter
encoding. DE encodes all the parameters using a floating
point, while classic GA uses a binary encoding scheme.
This real value encoding and the fact that DE parameters
Application
Estimation of Monod model parameters of a bioreactor
Estimation of kinetic parameters of the ethanol fermentation process
Feed batch fermentation
Feed batch fermentation
Feed batch fermentation
Optimal design of an ammonia synthesis reactor
Optimization of adiabatic styrene reactor
Estimation of heat transfer parameters in a trickle bed reactor
Optimal design of shell and tube heat exchanger
Heat exchanger network design
True boiling point curve of crude oil; effect of pressure on entropy
Estimation of optimal time of pyrolysis and heating rate
Alkylation process optimization
Alkylation process optimization
 E.N. Dragoi and S. Curteanu: Differential evolution in chemical engineering      151

are manipulated with arithmetic operators have several
advantages: efficient memory utilization, ease of use,
lower computational effort, and complexity (Price et  al.
2005). Along with efficiency, other advantages are flex-
ibility (the algorithm adapts to modifications) and funda-
mentality (the principle of differentiation synthesizing in
itself the fundamentals concepts of the solution search)
(­Feoktistov 2006).
2.1 Initialization
where bk,L and bk,U are the lower and upper limits of each
k characteristic and rand(0, 1) is a random generated
number in the interval (0, 1).
In a review about the initialization procedures for
large-scale optimization problems (Kazimipour et  al.
2013), the initialization techniques were classified into
(i) stochastic, (ii) deterministic, (iii) two steps, (iv) other
methods, and (v) application specific. These methods can
be also used in the DE case; the most employed variant
(except the deterministic approaches, which are the
classic choice) is the two-step variant, which includes
opposition-based learning (OBL) and quasi-OBL.

In this step, the initial population is created, usually by

employing a random-number-generating procedure, the
2.2 Mutation
algorithm selecting which locations should be explored
and which should be discarded (de Melo and Botazzo
In the mutation step, in order to create new individu-
Delbem 2012). Although it is an important step (as the per-
als, the selected individuals are modified by introducing
formance is affected by initialization and improper initial-
more genetic material into the population (Salman et al.
ization can lead to high errors), initialization is somewhat
2007). The role of mutation is to add diversity and to avoid
discarded, the majority of researchers mentioning in their
getting trapped into local optima. In the DE case, a new
works a series of general elements without any additional
individual is created by adding a scaled differential term
details (Ali et al. 2013).
to a base vector (individual) (Price et  al. 2005, Das and
In order to perform a proper initialization procedure,
Suganthan 2011). This mechanism [described in its simple
different aspects must be taken into consideration. The
form by Eq. (2)], also called “differentiation”, is the main
most important ones are represented by (i) the position
characteristic separating DE from other EAs:
of the global minimum (considering that the problem is
a minimization one) and (ii) boundary determination. ωi = α + F ⋅ β, (2)
Before solving an optimization problem, the position of
where ωi is the ith element from the mutated population, α is
the global minimum is usually not known, and in this
the base vector, β is the differential term, and F is the scaling
context, it is indicated to start the algorithm by using
factor. The differential term is defined as the difference of
equally dispersed points from the search space (Lei et al.
two distinct, randomly chosen vectors: β = xk-xp. The base
2010, Peng and Wang 2010). Some algorithms start from
vector is also randomly chosen, and in order to achieve good
a single point, but DE needs a population composed of
convergence speed and probability, Price et al. (2005) indi-
diverse individuals (Price et al. 2005).
cate that all vectors used in the mutation step must be dis-
Although there are problems that have unconstrained
tinct. This mechanism – differentiation – has the advantage
parameters, for the most real-world problems, the exist-
of not producing a disruption in the hyper-plan in which the
ence of natural physical limits or logical constraints
vectors exist (Feoktistov 2006).
imposes different values for each parameter, and their ini-
Initially, Storn (1996) presented 10 variants of the
tialization is a straightforward process. For example, the
algorithm, with a coding “Mode/DiffTerm/Cross” being
time for studying a phenomenon cannot be negative or
used to indicate each variant. “Mode” represents the
the pressure of a chemical reactor takes values from a pre-
mode in which the base vector of the mutation step is
specified interval, which is dependent on the characteris-
chosen, “­DiffTerm” represents the number of differential
tics of the reactor. In this context, initialization cannot be
terms used for mutation, and “Cross” is the type of crosso-
performed until all the boundaries are known.
ver. The first two terms (“Mode” and “DiffTerm”) refer to
Eq. (1) describes the formula used for initializing
the characteristics of the mutation step, and the last one,
each k characteristic of each i individual from the popula-
“Cross”, is related to the crossover step. The base vector
tion. This equation corresponds to the classic variant of
for the mutation step can be chosen randomly (Rand), as
DE, which, in the literature, has different variations and
the best individual in the population (Best), or as a vector
improvements.
that lies on a line between the target and the best individ-
xi ,k = bk ,L + rand( 0, 1) ⋅( bk ,U -bk ,L ), (1) ual (Rand-to-Best). Regarding the number of differential

Brought to you by | Universita degli Studi di Padova

Authenticated
Download Date | 3/28/16 10:48 AM
152      E.N. Dragoi and S. Curteanu: Differential evolution in chemical engineering
terms used in the mutation phase (information coded
with “DiffTerm”), one [Eq. (2)] or two terms are usually
employed, which are denoted with 1 and 2.
The first variant of the DE algorithm, which uses the
random method of vector selection, along with one differ-
ential term and binomial crossover, is called Rand/1/Bin.
The other versions included in the initial 10 are Rand/2/
Bin, Rand/1/Exp, Rand/2/Bin, Rand/2/Exp, Best/1/Bin,
Best/2/Bin, Best/1/Exp, Best/2/Exp, Rand-to-Best/1/Bin,
and Rand-to-Best/1/Exp.
Eq. (2) describes the variant in which only one dif-
ferential term is used. However, in literature, a series of
variants that use more differential terms can be encoun-
tered [Eqs. (3) and (4)]. Therefore, the search space is
k ∗2
extended, existing ∑ i=0 ( Np-i ) possibilities of combi-
nation of  2  *  k randomly chosen individuals (Feoktistov
2006). A high number of individuals leads to high diver-
sity, and ­Feoktistov (2006) recommends the use of three
to five individuals, which is a good compromise between
computational time and quality of exploration.
ui ,g = xi ,g + Fi ( xbest
p
,g
-xi,g ) + Fi ( xr 1,g -xr 2,g ), (3)
where xbest
p
,g
is a randomly chosen solution from the top
100 p% individuals from the g population, r1 and r2 are
*
two uniformly chosen indexes, and Fi is the scale factor
associated with the ith individual.
ui ,g = xr 1,g + Fi ( x pbest -xr 1,g + xr 2 ,g -xr 3,g ), (4)
where r1, r2, and r3 are three randomly and mutual dif-
ferent individuals from the g population and xpbest is a
randomly chosen individual from a Sp set generated at
the beginning of each generation according to a specific
procedure.
The variant described by Eq. (3) was proposed in
Zhang and Sanderson (2009), being called DE-current-to-
pbest/1. The one represented by Eq. (4) is an improvement
of the variant of Zhang and Sanderson (2009) and was
proposed by He and Yang (2012).
Except the number of differential terms, another
aspect that influences mutation in particular, and the
algorithm performance in general, is the scaling factor
F applied to the differential terms. Its role is to control
the rate at which the population evolves by determining
the perturbation size of the differential term. Usually,
the interval (0, 1) is considered for this parameter, but a
fixed upper limit was not determined. However, a limit of
F = 1.2 was empirically determined, optimization problems
requiring F to trespass this threshold not being encoun-
tered (Mallipeddi et al. 2011). In addition, there are only a
few optimization problems that need F > 1, a case in which
Brought to you by | Universita degli Studi di Padova
Authenticated
Download Date | 3/28/16 10:48 AM
the differential term is scaled up. A larger F can increase
(i) the probability of escaping the local optimum and
(ii) the number of function evaluations (Deng et al. 2013).
2.3 Crossover
In this step, an operation for population diversity
improvement is applied. Using two populations (current
and mutated), a new trial population is created. Generally,
two variants of crossover are used in the DE algorithm:
binomial [Eq. (5)] and exponential.
ω if ( rand( 0, 1) < Cr )
ui ,j =  i ,j , (5)
 xi ,j otherwise

where ui is the trial vector and xi is the ith individual from
the current population.
In the binomial version, the number of mutant com-
ponents used has a binomial distribution, the principle
behind it being simple: a random number in the (0, 1)
interval is generated and, if it is smaller than the crossover
control parameter (also known as Cr), then the element is
taken from the mutated individual. If the generated value
is higher than Cr, then the element from the current indi-
vidual is copied into the trial one. On the other hand, the
exponential crossover is somewhat different. An initial
start point (sp) is generated, and all the characteristics
with an index lower than sp are copied from the mutation
vector. When the index is higher than sp, a random value
between 0 and 1 is generated, and until it is bigger than Cr,
the characteristics from the current individual are copied
to the trial vector. After that, the remaining characteristics
are taken from the mutation vector.
The difference between the two mechanisms consists
in the position of characteristics taken from the parents.
In the binomial version, the components from the mutant
individual are randomly selected, while in the exponential
one they are organized into groups of compact sequences
(Zaharie 2002). Analysis of the differences between these
two variants showed that the use of exponential crossover
to the detriment of the binomial one increases efficiency
only for a small number of problems, and therefore,
researchers did not focus on the crossover as a mean to
improve performance (Tvrdik 2008).
A control parameter (Cr) is employed to control
which and how many components of each individual are
copied (Mallipeddi et  al. 2011). It can take values in the
[0, 1] interval, with its optimal value influenced by the
type of problem and type of crossover. For example, when
the objective function is nonseparable, a [0.9, 1] interval
 E.N. Dragoi and S. Curteanu: Differential evolution in chemical engineering      153
is considered suitable and a [0, 0.2] interval when it is
separable (Mohamed et al. 2013). Concerning the optimal
value of Cr based on crossover type, knowledge is limited
to the distribution of the number of mutated components
(Alguliev et al. 2012).
The classic crossover versions can generate only
one solution, limiting the search ability of DE (Wang
et  al. 2012). Consequently, new crossover variants were
proposed or borrowed from other algorithms. One such
example is represented by the orthogonal design method,
with different studies presenting the advantages of apply-
ing an orthogonal crossover for generating offspring
(Gong et al. 2006, 2008).
2.4 Selection and stop criteria
In the last step of the algorithm, a mechanism for selecting
the individuals forming the next generation is employed
(Rahnamayan and Tizhoosh 2008). The classic version of
DE uses a “one-to-one” competition, the trial and current
individuals being compared based on their objective
values. The ones with the lowest value (when consider-
ing a minimization problem) are selected to form the next
generation. This comparison determines in DE a tighter
integration of recombination and selection than in other
EAs (Price et al. 2005). This survival criterion is also called
“objective function based criterion” or “greedy criterion”.
In literature two other mechanisms can be encountered:
age criterion and age-objective function criterion (Price
et al. 2005, Li and Li 2010).
As the majority of problems have specific constraints,
special attention must be given to the parameters that
exceed the boundaries. When applying the selection cri-
terion, because DE does not have an innate mechanism
for dealing with constraints, a constraint handling mech-
anism must be used in order to evaluate, and if neces-
sary, to move the exceeding parameters back into their
feasible interval. Depending on the characteristics of the
problem being solved, different approaches can be used.
For example, in the case of single-objective problems,
the selection process is somewhat simple (as it is easy to
decide which between the two individuals is better), but
in the case of multioptimization problems, the decision is
not so straightforward (Qin et al. 2010).
The steps of the algorithm are repeated until a stop
criterion is reached. An optimal selection of this criterion
ensures a good tradeoff between solution performance
and consumed resources. In an ideal case, the optimiza-
tion is stopped when the convergence to the optimum is
achieved (Zielinski and Laur 2008). As the determination of
Brought to you by | Universita degli Studi di Padova
Authenticated
Download Date | 3/28/16 10:48 AM
this condition is tricky, this aspect usually cannot be used.
Therefore, researchers focused on finding a good solutions
for this problem, with the best variants proposed including
(i) the number of current generations reaching a predefined
maximum number (G) (as in Figure 1) or (ii) the number of
function evaluations reaching a maximum value (FEmax).
Figure 1 presents the general workflow of the DE algo-
rithm, in a simple/classic variant where the stop criterion
is represented by the number of generations reaching the
prespecified upper limit.
2.5 Algorithm improvement
In order to improve the algorithm performance and to
make it more flexible for the theoretical and real-life
applications, different modifications at various levels
have been performed. All these modifications followed
three main directions (Brest et al. 2011): (i) replacing the
hand tuning of control parameters with adaptive or self-
adaptive mechanisms, (ii) hybridizing DE by combining it
with other optimization techniques, and (iii) introducing
more mutation strategies during the optimization process.
Each control parameter influences specific aspects
of the algorithm, its effectiveness, efficiency, and robust-
ness being dependent on their correct values, which are
problem specific and can vary substantially (Brest 2009,
Hu and Yan 2009a). Due to their influences, finding the
optimal values of the control parameters is not a straight-
forward process, and the methods used for their correct
settings can be classified into parameter tuning and
parameter control (Eiben and Schut 2008). Parameter
tuning consists of finding good values before running the
algorithm. On the other hand, in the case of parameter
control, the values are changed dynamically during the
run based on a set of defined rules. Taking into account
the “how” criterion of Eiben and Schut (2008) indicating
how the change is performed, four subclasses are encoun-
tered: (i) deterministic control, where the parameters are
determined using a deterministic law, without feedback
from the system; (ii) adaptive control, where the param-
eters are determined using feedback from the system. Two
categories are encountered here: parameter adaptation,
which obeys the state of the population and refresh of
population; (iii) self-adaptive control, where the param-
eters are dependent of the algorithm being encoded into
it; and (iv) hybrid control.
The second direction of improvement – hybridization
– is the process of combining the best features of two or
more algorithms in order to create a new algorithm that is
expected to outperform the parents (Das and Suganthan
154      E.N. Dragoi and S. Curteanu: Differential evolution in chemical engineering

Figure 1: Schema of the simple DE variant.

2011). Depending on the type of algorithm DE can be
hybridized with, three situations encountered: (i) DE and
other global optimization algorithms, (ii) DE with local
search methods, and (iii) DE with global optimization and
local search methods. Depending on the level of interac-
tion, the hybridization can be performed at the (i) individ-
ual level, describing the behavior of an individual from the
population; (ii) population level; (iii) external level; and
(iv) meta-level, where a superior metaheuristics includes
the algorithm as one of its strategies (Feoktistov 2006).
In case of the third direction of improvement, differ-
ent approaches related to the mutation were employed.
The classic variant regarding mutation in DE is to create,
for each individual in the population, a single mutated
individual based on one or more vectors. The alterations
proposed by researchers refer to the (i) introduction of
more differential terms, (ii) selection of the individuals
participating in the mutation process using specific prin-
ciples, (iii) introduction of other mechanisms (evolution-
ary game theory, Pant et  al. 2009a, and replacement of
control parameter with Laplace distribution, Thangaraj
et  al. 2010), and (iv) employment of multiple mutation
strategies (Qin et al. 2009, Zamuda and Brest 2012).
The principles from all these directions can be com-
bined, with DE being modified at different levels, using dif-
ferent principles. For example, opposition-based chaotic
DE algorithm (Thangaraj et al. 2012) has an OBL-based ini-
tialization and the crossover factor is adapted using chaos
theory; DE-CCS (Wang and Gao 2014) (DE with coopera-
tive coevolution) has a jumping DE (jDE)-based approach
for the control parameters, a crossover scheme that
employs both binomial and exponential crossover and a
local search using cooperative coevolution; in hybrid self-
adaptive DE with neural networks (Curteanu et al. 2014),
neural networks are optimized using a DE-based version
that includes OBL initialization, a modified mutation and
a local search procedure based on BackPropagation and
Local Search algorithms.
3 Process optimization
Process optimization refers to the action of adjusting dif-
ferent aspects of the process in order to make it more effi-
cient and to minimize costs. There are three areas in which
the process optimization can be performed: equipment,

Brought to you by | Universita degli Studi di Padova

Authenticated
Download Date | 3/28/16 10:48 AM
 E.N. Dragoi and S. Curteanu: Differential evolution in chemical engineering      155
operating procedures, and control optimization. In order
to perform such a task, a series of mathematical systems
describing the process is necessary. As, often, control opti-
mization is closely related to the equipment and operation
procedures, and the boundaries between all these areas
are somewhat blurry, all the works related to the process
optimization will be presented in the same section.
3.1 S
 imple DE versions used in process
optimization
In this section, different optimization procedures based
on the classic DE versions (without any improvements
such as adaptation or hybridization) applied to specific
chemical engineering processes are presented.
In Moonchai et al. (2005), the effect of pH and temper-
ature on the bacteriocin production in Lactococcus lactic
C7 was studied and a series of models for five variables
were developed. In order to determine the optimal temper-
ature profile (which, in this case, was considered a control
variable), the temperature parameter was introduced in
the models created for cell growth, lactic acid production,
glucose consumption, and bacteriocin production. After
that, using a weighted sum method, the two objective
problems (determination of temperature control leading
to the maximization of productivity and final bacteriocin
titer) were converted into a single objective problem. It
was then solved using the DE algorithm, the simulation
results indicating that the bacteriocin production rate and
titer are at the highest level when starting with a 30°C tem-
perature and then lowering it quickly at 22°C.
As the dynamic optimization of the fermentation
processes is a problem requiring not only powerful algo-
rithms but also carefully chosen methodologies, Rocha
et  al. (2007) proposed an online optimization procedure
based on DE. Three different case studies were consid-
ered: (1) fed-batch recombinant Escherichia coli fermen-
tation, where microorganisms follow three different
metabolic paths (oxidative growth on glucose, fermenta-
tive growth on glucose, and oxidative growth on acetic
acid); (2) a hybridoma reactor; and (3) fed-batch bioreac-
tor for the production of ethanol by Saccharomyces cer-
evisiae. In case 1, the scope was to determine the optimal
feeding rate profile maximizing productivity (which was
represented by the recombinant protein formed per unit of
time). For case 2, the objective was to increase the amount
of monoclonal antibodies produced by the system, and in
case 3, the aim was to determine the substrate feed profile
leading to the maximization of ethanol amount. If in case
of the offline optimization small noise leads to disruption
Brought to you by | Universita degli Studi di Padova
Authenticated
Download Date | 3/28/16 10:48 AM
in the predictions, the online approach exhibits properties
of graceful degradation. However, when noisier settings
were employed, the DE algorithm performance decreased.
Using the same case studies and objectives (from Rocha
et  al. 2007), Mendes et  al. (2008) performed online and
offline optimizations of input variables by applying three
distinct algorithms, namely DE, a real value EA, and a
fully informed particle swarm (FIPS). The results obtained
for the offline optimization indicated that DE/Rand/1/
bin has the best results, followed closely by DE/best/2/
bin and FIPS. Similar findings were obtained when online
optimization was performed.
Khademi et  al. (2009) studied the optimization of
methanol synthesis and cyclohexane dehydrogenation
in a thermally coupled reactor by means of a Best/1/bin
strategy with fixed control parameters (Cr = 1, F = 0.8, and
Np = 100, with Np being the number of individuals in the
population). Their main objective was to maximize the
methanol and benzene mole fractions based on three
parameters: inlet temperature of the exothermic and
endothermic sides and initial molar flow rate of endo-
thermic side. It was shown that suitable values of the
considered parameters can provide the necessary heat
to drive the endothermic process and, in the same time,
to enhance the methanol mole fraction with 3.67% and
methanol yield with 3.1%. In another study (Khademi
et  al. 2010), the same research group proposed a new
membrane thermally coupled reactor for methanol syn-
thesis, cyclohexane dehydrogenation, and hydrogen pro-
duction. The new system was optimized with the same DE
version, having the same control parameter values, the
scope being the maximization of outlet mole fractions of
methanol, benzene, and hydrogen. Six decision variables
were chosen: inlet temperature and molar flow rate of the
endothermic and exothermic sides, inlet pressure of the
exothermic side, and inlet temperature of the permeation
side. Compared with the conventional methanol reactor
(CMR), the optimized system had a reduced size of reactor,
enhanced equilibrium conversion, and an increased
methanol production with 16.3%.
A novel methanol synthesis loop with hydrogen
permselective membrane reactor was modeled and opti-
mized using a Rand/2/bin version (Parvasi et  al. 2009).
The behavior of the system was described by a set of ordi-
nary differential and algebraic equations, which was then
used as a model for the optimization procedure with the
final scope of improving methanol production based on
the inlet temperatures of the membrane tube and steam
drum. The new design had a 40% increase in production,
which indicated that the applied approach was suitable
for the considered process.
156      E.N. Dragoi and S. Curteanu: Differential evolution in chemical engineering
Another type of reactor optimized with DE was rep-
resented by a radial-flow, spherical bed methanol syn-
thesis reactor in the presence of catalyst deactivation
(­Rahimpour et al. 2009). The Rand/2/bin version was used
in four cases for maximizing the methanol production
based on (i) the inlet temperature of reactors, (ii) tempera-
ture profile, (iii) radius ratios of reactors, and (iv) catalyst
deactivation parameters. In each case, the optimization
was successful, with the considered parameters being
determined with great accuracy.
Catalytic naphtha reforming is a technology of utmost
importance for the petrochemical industry, with effort
and energy being spent on improving or replacing the
classic approaches and systems such as the conventional
tubular reactor (CTR) with newer and improved reactors.
The dynamic of a multistage spherical, radial flow reactor
for the naphtha reforming process in the presence of cat-
alyst deactivation was performed using a DE approach
(Best/1/bin) (Rahimpour et al. 2010). Based on seven deci-
sion variables (inlet temperature and catalyst distribution
of each reactor, outlet gas pressure), the maximization
of production of aromatics and hydrogen was followed.
The results indicated that by using higher temperature
furnace (allowing gas to enter at 840 K), the efficacy of the
system rises.
In another study, the CTR was optimized via DE
method in three cases and the reactor was discretized
in 20 segments (Iranshahi et  al. 2011). The optimiza-
tion objective was the maximization of hydrogen and
aromatics productions based on (i) temperature profile,
(ii) amount of hydrogen removed, and (iii) a combination
of temperature profile and hydrogen removal. The best
results were obtained in the third case, an improvement
of 24% for aromatics and 10% for hydrogen production
being predicted.
The electrical parameters (six temperature-depend-
ent and current-dependent nonlinear resistors) of a biohy-
drogen real-time power-generating system were optimized
using a DE version with fixed control parameters (F = 0.8,
Np = 300) (Huang et  al. 2010). The proposed proton
exchange membrane fuel cell system (with a rated power
of 3 W) had bipolar plated designed for low-pressured bio-
hydrogen. Using the optimization results, the V-I charac-
teristics were determined based on the Kirchhoff’s law. In
addition, a comparison between the DE predictions and
those generated by GA was performed, with the optimal
temperature estimated by DE being 336.89  K and by GA
being 303 K.
Using a multi-input, multioutput orthogonal least-
square-based radial basis function model and a classic
DE approach with fixed parameters (F = 0.5, Cr = 0.9, and
Brought to you by | Universita degli Studi di Padova
Authenticated
Download Date | 3/28/16 10:48 AM
Np = 100), the production process of para-diethylbenzene
through disproportionation of ethylbenzene was modeled
and optimized (Rahman et  al. 2010). The selectivity and
conversion of para-diethylbenzene were modeled as a
function of temperature, liquid hourly space velocity, and
pressure.
Dimethyl ether (DME) is another fuel that has cap-
tured the attention of researchers, as it is a clean multi-
purpose and multisource fuel (Vakili et  al. 2011). In the
work of Vakili et  al. (2011), the operation conditions of
a thermally coupled direct DME heat exchanger (TCDR)
were optimized using a Best/1/bin version of DE with
Cr = 0.8, F = 0.8, and Np = 40. The scope was to maximize
the DME and benzene production based on a set of deci-
sion variables represented by the inlet temperature and
initial molar flow rates of the exothermic and endother-
mic sides. Compared with the conventional DME synthe-
sis reactor (CDR), the optimized system had a reduced
inlet feed of exothermic side with about 4300 kmol/
day and a 1.6 ton/day increased DME production. Con-
cerning the endothermic side, the inlet feed is reduced
with 25,700  kmol/day and the hydrogen production is
increased with 0.8 ton/day.
In another work, the DME synthesis and cyclohex-
ane dehydrogenation reaction taking place in a thermally
coupled reactor were optimized using classic Best/1/bin
version (Khademi et al. 2011). The optimization goal was
to determine the operating conditions (temperature pro-
files, endothermic and exothermic sides) that maximize
the DME and benzene mole fractions. The decision vari-
ables considered were inlet temperatures of the exother-
mic and endothermic sides and initial flow rates of the
endothermic and exothermic sides. The final result had an
improvement of 2.43% for the DME fraction and of 2.78%
for methanol conversion.
Rahimpour et al. (2011) used the DE algorithm for opti-
mizing the operating conditions of a thermally coupled
membrane reactor for decalin dehydrogenation and FT
synthesis in gas-to-liquid (GTL) technology. The gaso-
line yield (exothermic side) and hydrogen model fraction
(endothermic side) were maximized and hydrogen molar
flow rate (recycled stream) was minimized based on five
decision variables (inlet temperature and flow rate of the
endothermic and exothermic sides and exothermic pres-
sure). The Best/1/bin variant with fixed control parame-
ters (F = 1, Cr = 0.8, and Np = 100) had good results, with an
increase of 14.28% of gasoline yield.
A fixed-bed tri-reformer reactor (combining endother-
mic reforming reaction and exothermic oxidation reac-
tion) for methanol production was optimized in order to
determine the optimal operating condition leading to a
 E.N. Dragoi and S. Curteanu: Differential evolution in chemical engineering      157

maximization of methane conversion, hydrogen produc-
tion, and desired H2/CO ratio (Arab Aboosadi et al. 2011).
The classic DE-based optimization procedure with penalty
function used eight decision variables: inlet composition
of CH4, CO2, CO, H2O, H2, O2, and N2 and inlet tempera-
ture of gas feed. The optimized system had an increase
in methane conversion from 94.3% to 97.9%, a H2 yield of
1.84%, and a H2/CO ratio of 1.7%.
The estimation of the biomass concentrations at a
product scale feed batch process was studied by five dif-
ferent estimation methods: (i) kinetic model of overflow
metabolism, (ii) metabolic black-box model, (iii) observer,
(iv) artificial neural networks (ANNs), and (v) an adap-
tive DE (RandToBest/1/bin version) (Hocalar et  al. 2011).
Two different technical scale bubble column fermenters
represent the system modeled, where S. cerevisiae yeast is
used as host organisms for recombinant proteins and for
producing baker yeast and ethanol. In order to evaluate
the performance of the selected approaches, four types of
fermentation were considered. The algorithms were com-
pared in terms of difficulty of implementation, number of
parameters measured, and process knowledge. The mean-
squared errors (MSE) obtained with all “fermentation
type-modeling” combinations are presented in Table  2.
The results pointed out that each procedure has advan-
tages and disadvantages, in the DE case, a few measure-
ments being required and the convergence being good for
different fermentations. On the other hand, DE has a long
response time that depends on the control parameters.
In this case, one possible improvement, in both response
time and efficiency, is represented by the use of a DE self-
adaptive procedure in which the control parameters (espe-
cially the crossover probability and the scaling factor) are
internally adapted to the specific characteristics of the
search space problem.
In a review related to the use of EAs for optimization
of a fermentation process, Adeyemo and Enitan (2011) list
a series of applications of the DE algorithm. What can be
observed is that GAs and ANNs are widely used and the
number of works is considerable, with these algorithms
establishing themselves as “to go tools” when specific
aspects cannot be solved with classic approaches. On the
other hand, the application of DE for fermentation pro-
cesses is still in its infancy.
In the study of Vakili et al. (2012), a new DE optimized
thermally coupled membrane configuration (TCMDR)
simultaneously producing hydrogen and DME was pro-
posed. This system was based on the TCDR developed in
Vakili et al. (2011) to which a membrane concept is added.
The main role of the optimization procedure (represented
by the Best/1/bin version with fixed control parameters
Np = 50, F = 0.8, and Cr = 0.8) was to identify the optimum
design of the configuration and maximization of hydro-
gen recovery and of production yield (DME, benzene).
Compared with the CDR and TCDR, the DME production of
the optimized TCMDR was improved by 10.3% and 11.4%,
respectively. Also, the cyclohexane conversion was raised
by 8.3% (compared with TCDR) and the endothermic flow
rate was reduced, with the operation cost of the system
being lower.
Another type of reactor proposed for direct DME pro-
duction is represented by the industrial scale dual-type
reactor, combining two fixed-bed reactors (one water
cooled and one gas cooled) (Vakili and Eslamloueyan
2012). In order to achieve the best performance, the reactor
design was combined with an optimization procedure,
with a set of seven decision variables (number of tubes
in the water and gas cooled reactors, temperature of feed
gas, reactor length, coolant temperature water in the first
reactor, second reactor diameter) being considered. The
optimization procedure employed a Best/1/bin version
with Np = 70, F = 0.8, and Cr = 0.8. Compared with CDR,
the optimized proposed system had an increased DME
production, with 60 ton/day. In addition, by eliminating

Table 2: MSEs obtained with different estimation methods in the case of four fermentation types of Saccharomyces cerevisiae.

Fermentation   Estimation method

 
Kinetic model of   Metabolic black-   ANNs   Observer   DE
overflow metabolism box model

Fermentation producing minimum ethanol, in the 100 m3   0.1848   0.1296   0.2638   0.3777   0.1213
reactor
Fermentation producing and then consuming a small amount of  0.4108   0.0697   0.2608   0.3580   0.2107
ethanol in the first hour of fermentation, in the 100 m3 reactor
Fermentation producing around 1% ethanol during the   0.3315   0.0875   0.2964   0.3328   0.1725
exponential phase, in the 100 m3 reactor
Fermentation producing and consuming around 0.2% ethanol   0.2597   0.2181   0.4311   0.6145   0.2010
in the first hours of fermentation, in the 25 m3 reactor

Brought to you by | Universita degli Studi di Padova

Authenticated
Download Date | 3/28/16 10:48 AM
158      E.N. Dragoi and S. Curteanu: Differential evolution in chemical engineering
the need for a separate unit for methanol production and
purification, the production costs were reduced.
In Arabpour et al. (2012), in order to enhance the gas-
oline production, the optimization of the Fischer-Tropsch
(FT) synthesis reactions in GTL technology was performed
using Best/1/bin version in case of five discretized reactor
models. For each model, different aspects were followed:
temperature optimization (first model, D1), optimization of
injected hydrogen (second model, D2), effect of optimum
removed water (third  model, D3), effect of temperature
and hydrogen removal (fourth model, D4), and simultane-
ous optimization of temperature, injected hydrogen, and
removed water (fifth model, D5). The reactor length was
divided into 20 elements, with the optimized values for
each model and for each element being determined by DE,
starting from the first segment from the top of the reactor.
The results indicated that the highest yield of gasoline is
obtained in the D5 model. Compared to the conventional
reactor, in D1 to D5 cases, the CO2 production is reduced by
6.2%, 21.6%, 27.6%, 24.1%, and 31.4%, respectively.
In Rahimpour et  al. (2012), a process intensification
procedure of the catalytic naphtha reforming unit was
performed by optimizing a system based on a thermally
coupled naphtha reforming reactor (TCR). The scope was
to maximize the aromatic hydrogen and aniline molar
flow rates and the aromatic in reformate mole fractions by
using several decision variables, from which 10 belong to
the endothermic side (naphtha reforming) and 16 belong to
the exothermic side (nitrobenzene hydrogenation). Com-
pared with the nonoptimized TCR, the optimized system
has a 26.3% increase in aromatic yield (endothermic side)
and a 39.21% increase in aniline production (exothermic
side). The higher thermal efficiency and lower operational
cost of the optimized system indicated that the proposed
approach is a good way to boost production.
In another study, the operating conditions of a radial-
flow tubular membrane reactor (RF-TMR) are optimized
(Iranshahi et al. 2012). As in the previously presented study
(Rahimpour et al. 2012), a relatively high number of param-
eters were used for optimization. Some of them are specific
to the first, second, and third reactors (sweep gas pressure,
membrane thickness, inner radius flow of the collector,
catalyst mass distribution, fraction of the total sweep gas,
number of sweep sections, ratio of permeation angle, and
ratio of length per diameter), whereas some are specific to
the sweep gas (total molar flow and hydrogen molar frac-
tion), recycle gas (hydrogen mole fraction), and to the first
reactor (total fresh naphtha feed). Compared with the con-
ventional, dimensionless tubular reactor, the optimized
radial flow tubular membrane reactor had an increase of
11.6% for hydrogen and 8.9% for aromatic yields.
Brought to you by | Universita degli Studi di Padova
Authenticated
Download Date | 3/28/16 10:48 AM
The operational conditions in both sides of a cata-
lytic thermally coupled fluidized bed naphtha reactor
(TCFBNR) were optimized by using a DE technique
(Pourazadi et al. 2013). Based on a penalty function (with
a parameter equal to 108), a set of constraints guarantee-
ing an increase in aromatic and hydrogen production
were imposed. The results obtained indicated that the
number of tubes can be decreased by 29% compared to
the non-optimized TCFBNR. Also, through optimization,
an increase in aromatic production (464.4 kg/h), hydrogen
production (3.03  kg/h), and aniline flow rate (14%) was
realized.
Iranshahi et al. (2013) proposed an alternative to the
CTR by developing a combination of tubular membrane
and radial-flow spherical reactors (TM-RSR). In addition,
the new system was optimized by DE, the optimal design
parameter, and the operating conditions being used. Two
configurations, combining a radial flow spherical reactor
with one tubular membrane reactor (case I) or with two
tubular membrane reactors (case II), were considered.
The number of parameters and their meaning vary in both
cases, with a list of all the parameters used for optimiza-
tion in this work (Iranshahi et al. 2013) and the ones pre-
viously discussed (Iranshahi et al. 2012, Rahimpour et al.
2012, Pourazadi et  al. 2013) being presented in Table 3.
The configurations used for case I were MSS, SMS, SSM,
and SSS, and the configurations for case II were MMS,
MSM, SMM, and MMM, where S denotes the spherical
reactor and M denotes the tubular membrane. The best
variants were SMS and SMM, with their optimized variant
(compared with CTR) having a considerable increase in
the hydrogen (23 kmol/h) and aromatic (6.5 kmol/h) pro-
duction rates.
Samimi et al. (2013a) proposed a DE optimized ther-
mally coupled reactor with a dual membrane (TCDMR) for
simultaneous production of DME, hydrogen, and naph-
thalene. After determining the system models using the
kinetic equations, energy, and mass balances in addition
with several auxiliary correlations, the operation con-
ditions of the thermally coupled reactor (case 1) and of
TCDMR (case 2) were optimized using a Best/1/bin version
with fixed control parameters (Np = 100, F = 0.8, and Cr = 1).
In the first case, the scope was to maximize the DME mole
fractions on the outlet exothermic side based on three
decision variables (inlet temperature, initial molar flow
rate, and inlet pressure of the exothermic side). In the
second case, the maximization of the outlet exothermic
DME mole fraction and endothermic hydrogen mole frac-
tion was considered, eight decision variables being used
(inlet temperature of water permeation, hydrogen perme-
ation, endothermic and exothermic sides, inlet pressure,
 E.N. Dragoi and S. Curteanu: Differential evolution in chemical engineering      159

Table 3: Parameters optimized for catalytic naphtha reactors using the DE algorithm.

Parameter/Study   Reactors  TCR  RF-TMR  TCFBNR  TM-RSR

  (Rahimpour et al. 2012) (Iranshahi (Pourazadi  (Iranshahi et al. 2013)
et al. 2012) et al. 2013)
Endoth. side  Exoth. side Case I  Case II

Length to diameter  
Sweeping gas pressure  
Hydraulic diameter  
Membrane thickness  
Catalyst mass distribution  
Inner radius  
The fraction of the total sweep gas  
Number of sweep sections  
The ratio of permeation angle to the  
reaction one’s
Exothermic side molar flow rate  
Inlet pressure of the exothermic side  
Inlet molar flow rate of exothermic side  
Number of tubes  
Naphtha feed pressure  
Sweeping gas molar flow rate  
Hydrogen mole fraction in recycled reactor     *    *    *  *
Fresh naphtha feed molar flow rate  
Hydrogen mole fraction in sweeping gas  
Naphtha inlet temperature  
Sweep gas inlet temperature  
Compressor discharge pressure  
Total molar flow rate for the exothermic side      *
Inlet temperature of endothermic side  
1, 2, 3  *  *  *    * 3  *
1, 2, 3      *    * 3  *
1, 2, 3          * 3  *
1, 2, 3      *    * 4  *
1, 2, 3  *  *  *    *  *
1, 2, 3      *    *  *5
1, 2, 3      *     
1, 2, 3      *     
1, 2, 3      *     
1, 2, 3    *       
1, 2, 3    *    * 2   
1, 2        *   
1, 2, 3    *    * 2   
1  * 1    *    *  *
      *    *  *
          *  *
      *    *  *
  *        *  *
          *  *
1  *    *     
       
        *   

*1 Total naphtha feed.

*2 For reactors 1 and 2.
*3 Parameters of tubular reactor.
*4 Membrane thickness (not applied to the three reactors).
*5 Inner radius of the spherical reactor.

and feed rates of the endothermic and exothermic sides).
Compared with CDR, for DME mole fraction, a 3.42%
(case 1) and a 34.2% (case 2) improvement was observed.
Concerning the hydrogen mole fraction, a 6.4% improve-
ment of case 2 over case 1 was realized.
The operating conditions and the length per radius of
a novel axial flow spherical packed bed membrane reactor
for DME production through methanol dehydration were
optimized using the Best/1/bin version (Samimi et  al.
2013b). The scope was to maximize the DME mole frac-
tions in the reactor outlet, using a set of 11 variables: inlet
temperatures, molar flow rates, composition of compo-
nents in the reaction and permeation sides, reaction side
pressure, and length per radius. The kinetic model had a
2.15% maximum absolute error and the optimization leads
to a 13.5% increase compared with CDR.
In Samimi et  al. (2013c), a novel thermally double
coupled reactor is proposed for hydrogen, methanol,
and high-octane gasoline production. The system (a
multitubular reactor where two exothermic reactions rep-
resented by methanol production and FT synthesis provide
the heat required for the endothermic decalin dehydroge-
nation) is optimized using Best/1/bin variant with fixed
control parameters (F = 0.8, Cr = 1 and Np = 100). The objec-
tive was to maximize the summation of the methanol mole
fraction (endothermic) and outlet gasoline yield (exother-
mic) based on five decision variables: inlet temperature
and inlet pressure at the FT and methanol synthesis sides
and inlet flow rate at the FT side. Compared with CMR, the
methanol production rate was increased by 10.52%.
In Rahimpour and Mirvakili (2013), a decalin and
hydrogen looping approach named thermally coupled
membrane ternary reactor (TCMTR) for application in GTL
technology is proposed. In order to minimize the H2 mole
fraction in the outlet of the third reactor and to maximize
the gasoline yield and hydrogen mole fraction in the endo-
thermic side, the operating conditions of the TCMTR are
optimized using a DE approach. A set of seven decision

Brought to you by | Universita degli Studi di Padova

Authenticated
Download Date | 3/28/16 10:48 AM
160      E.N. Dragoi and S. Curteanu: Differential evolution in chemical engineering
variables were considered, including inlet temperatures
of the endothermic and exothermic sides, inlet molar flow
and pressure of the exothermic side, initial flow rate of
the endothermic side, temperature of the shell side, and
length of the third reactor. The results obtained indicated
that in the new proposed configuration, decalin can be
used as a hydrogen carrier, with only a small amount of
this substance being required for injection in the first
reactor. Also, compared with a thermally coupled mem-
brane dual-type reactor, the DE-optimized TCMTR has a
92% decrease in decalin consumption.
In Bayat et  al. (2013), the industrial fixed-bed ethyl-
ene oxide reactor was optimized (at both steady-state and
dynamic conditions and taking into account the cata-
lyst deactivation) using DE algorithm. Distinctively from
the Lahiri et al. (2008) work, a mathematical model was
employed, and two different cases were investigated with
the scope of maximizing the ethylene oxide production
rate during 1100 days of catalyst life. In the first case, four
parameter were optimized: inlet pressure of the reaction
side, inlet feed gas temperature, inlet cooling saturated
water, and inlet molar flow rate. For the second case
study, six parameters were optimized, from which three
were related to the inlet gas temperature and three to the
cooling water temperature. Based on the results obtained
in the first case, in the second situation, the optimal
behavior of cooling water and feed gas temperature were
determined. In the end, an improvement of 1.726% (first
case) and 4.22% (second case) was obtained.
A design method for simultaneous optimal determi-
nation of process configuration, equipment design, and
operating conditions of a membrane-assisted separation
process was developed by Koch et al. (2013). The consid-
ered case study was represented by separation through
distillation and membrane separation of the ternary
system acetone/isopropyl alcohol/water. The steps of the
new approach consisted of (i) generation of process alter-
natives through thermodynamic analysis, (ii) experimen-
tal investigations in a laboratory-scale membrane plant,
(iii) generic modeling based on mass-transfer models,
and (iv) process superstructure optimization with DE. The
variables optimized include distribution feed, distribution
side, distribution recycle, number of membrane modules
in series, number of membrane sheets in a module, total
weight of active packing for sections 1 to 4, reflux ratio,
flow rate of side stream, and molar recycle ratio. The scope
was to reduce costs, with the objective function contain-
ing investments and operation costs.
For methanol synthesis, a dynamic mathematical
model of a membrane-gas-flowing solid-fixed-bed reactor
with in situ water absorption, in the presence of catalyst
Brought to you by | Universita degli Studi di Padova
Authenticated
Download Date | 3/28/16 10:48 AM
deactivation, was proposed in the work of Bayat et  al.
(2014). In addition, an evaluation of the optimal operating
conditions was performed using DE, the results indicating
that under the optimal values of gas phase inlet tempera-
tures, flowing solids phase, and shell side, a high meth-
anol production can be achieved. A very similar study
related to the same type of reactor and reaction was also
performed in Dehghani et al. (2014).
As can be observed, different aspects of problems
belonging to the chemical engineering area can be solved
effectively with simple DE variants. The problems range
from fermentation to catalytic naphta reforming to fuel
and biofuel production. For the same process, various
aspects and conditions are studied from multiple points of
view. In this manner, the findings of different studies can
be combined, and useful conclusions for the experimental
practice derived.
Concerning the DE algorithm, the researchers did not
focus only on a single variant, the literature providing
examples where not only the type of mutation or crossover
were changed, but also the selected values for the control
parameters. As each process has its own characteristics,
which change even when different parameters are taken
into consideration, and as DE successfully solved all
aspects, it can be concluded that even in the simplest
forms, the algorithm is flexible enough to provide satis-
factory solutions.
3.2 M
 odified and hybrid DE variants in
process optimization
Kapadi and Gudi (2004) extended the optimal feed policy
of a fed-batch fermentation process proposed in Chiou and
Wang (1999) by including (i) systems with multiple feeds,
(ii) nonuniform parameterization (the time duration of
each interval is not the same), and (iii) path and end-
point constraints. In order to demonstrate the effective-
ness of the methodology, the same case study as the one
from Chiou and Wang (1999) was employed. In addition,
the dynamic optimization of simultaneous saccharifica-
tion and fermentation of starch to lactic acid was studied.
The goal was to maximize the final concentration of lactic
acid, productivity, and yield. In the case of a single feed
approach, DE in combination with the Lagrange-like
method, including dynamic penalty functions, provided
better results than Pontryagin’s maximum principle.
When using the nonuniform control vector parameteriza-
tion, the optimization was performed twice, first setting
the fermentation time as a decision variable, and second,
considering it fixed at 80 h. The results of the optimization
 E.N. Dragoi and S. Curteanu: Differential evolution in chemical engineering      161
problem (with 79 decision variables, from which 40 were
constant feed rates and 39 were time interval parameters)
indicated that nonuniform control vector parameteriza-
tion and endpoint constraints used in DE are more effi-
cient for end-product concentration and production rate.
Good results were also obtained in case of multiple feed
optimization.
In the work of Babu et al. (2005), a multioptimization
DE version (MODE) was employed to optimize the styrene
production by dehydrogenation of ethyl benzene in an adi-
abatic reactor. Five case studies corresponding to different
combinations of objectives were studied based on ethyl
benzene feed temperature, pressure, steam over reactant
ratio, and initial flow rate of ethyl benzene. Maximization
of styrene selectivity (SST) and styrene productivity (FST)
represents the first case, maximization of SST and styrene
yield (YST) represents the second case study, and maximiza-
tion of FST and YST is the third case study. In the fourth case,
the maximization of FST and minimization of H2O produc-
tivity are considered, while the last case is represented by
maximization of YST, FST, and SST. The same MODE approach
proposed in Babu et  al. (2005) was applied in another
study for optimization of two styrene reactor configura-
tions (single bed adiabatic operation and stream injected
pseudo-isothermal operation) (Gujarathi and Babu 2010a).
Five decision variables were considered for the adiabatic
operation (feed temperature, pressure, steam over reac-
tant ratio, initial ethyl benzene flow rate, and steam tem-
perature), while for the steam injected reactor, two more
are added (steam fraction at the reactor inlet and fraction
between the reactor length and location of the injection
port). Four optimization cases were considered, including
(i) maximization of FST and SST, (ii) maximization of SST and
YST, (iii) maximization of FST and YST, and (iv) maximization
of FST, SST, and YST. For both configurations, in all four cases,
Pareto optimal sets were obtained.
A simple self-adaptive approach (utilizing a co-evo-
lutionary proves running on a fractional time scale with
respect to the main process and working similarly to a high-
level co-evolutionary strategy) was applied for designing
a block decentralized PI controller of an ALSTOM gasi-
fier (Nobakhti and Wang 2008). The control system was
required to minimize the integral absolute error of the fuel
gas pressure and caloric value even when disturbances
appear (a 0.2 bar step reduction in the sink pressure and
a sinusoidal disturbance of 0.2 bar amplitude at 0.04 Hz).
These requirements were formulated into an optimization
problem with 20 design parameters and 26 constraints,
with the DE optimization providing near-optimum results,
a noticeable improvement over previously reported solu-
tions being observed.
Brought to you by | Universita degli Studi di Padova
Authenticated
Download Date | 3/28/16 10:48 AM
The catalytic industrial ethylene oxide reactor was
modeled and optimized using a DE-ANN approach (Lahiri
et  al. 2008). As a rigorous mathematical model of the
considered process (quantifying all the effects of param-
eters on selectivity) does not exist, the role of the neural
network was to act as a model, correlating all the process
data and performance variables. DE determined the
process input variables (which were also the model input
variables) leading to a maximization of yield and selec-
tivity. The considered inputs of the model were cycle gas
flow and pressure; reactor temperature; reactor inlet con-
centration of ethylene, CO2, oxygen, and ethane; oxygen
flow in the reactor; inhibitor flow; and catalyst running
hours, while the outputs were ethylene oxide production,
selectivity, and CO2 production. These parameters were
chosen based on a set of criteria, which included reduced
input number leading to the best prediction, high corre-
lation of input-output, and minimum cross-correlation
between inputs and minimum of network complexity.
After determining the model, a simple DE version (F = 0.5,
Cr = 0.2, and Np = 100) was successfully used to optimize
the model inputs corresponding to the set of process
operation conditions. The optimized solutions resulted
in a significant improvement in the ethylene oxide pro-
duction rate and catalyst selectivity. The DE optimization
program can be integrated with an online, real-time APC
application, or the optimum value may be calculated on
an off-line c­ omputer and further used for optimizing the
reactor.
Three DE-based versions were employed to optimize
the liquid phase oxidation of p-xylene to purified there-
phthalic acid in four different cases (Gujarathi and Babu
2009). The scope of this work was to maximize the total
feed rate (first objective) and to minimize the concentra-
tion of the 4-carboxybenzoic acid as intermediary product
(second objective). The first objective plays as a deci-
sion variable along with (i) catalyst concentration (CO),
(ii) CO and water content in the solvent (WII O), (iii) CO,
2
(WII O), unconverted reactant (PX), and vent oxygen from
2
the reactor and crystallizer merged into one variable VO , 2
and (iv) CO, WH O, PX, VO , and temperature of the reactor.
2 2
The DE versions used are represented by MODE (a variant
proposed in Babu et al. 2005); a modified MODE approach
called MODE-II, in which, at each generation, the popu-
lation is randomly generated (Babu 2007); and an elitist
MODE version (E-MODE), in which the concept of elitism
with crowded distance sorting is included. The results
showed that E-MODE provided widely spread, equally
spaced Pareto fronts, while MODE covered a wide area
of Pareto fronts in detriment of nonuniformly distributed
solutions.
162      E.N. Dragoi and S. Curteanu: Differential evolution in chemical engineering

The same process was also studied by Xu et al. (2013),
a self-adaptive DE strategy for solving constrained mul-
tiobjective problems proposed in Qian et al. (2012) being
used to improve the product yield. A set of conflicting
objectives (minimization of combustion loss and maximi-
zation of therephthalic acid yield) were considered, based
on nine process variables (concentrations of cobalt and
manganese catalysts, concentration of bromide promoter,
and temperature and residence time of the three reactors).
The system model was developed in Aspenplus 2006.5,
and the DE-based algorithm (with two versions, Rand/1
and Best/1) was implemented in Matlab, with an interface
between the two environments being created to perform
the optimization. In order to determine the performance
of this approach, a comparison with NSGA-II and fuzzy
immune algorithm was realized, with the results indicat-
ing that DE is effective in solving the considered problem.
A hybrid strategy combining support vector regression
(SVR) and DE was used to model and optimize ultravio-
let (UV) transmittance of the monoethylene glycol (MEG)
product of a commercial petrochemical plant (Lahiri and
Khalfe 2010). The procedure was designed to maximize
the UV based on a set of process parameters (reflux ratio
and flow, MEG column top and condenser pressure, MEG
column control temperature and feed flow, drying column
control and bottom temperature, and off-spec glycol
reprocessing flow). After the SVR model was determined,
DE was employed to optimize the N-dimensional model
input in three temperature case limits, which were then
verified in an actual plant.
In Yuzgec (2010), the feeding profile of an industrial
scale baker’s yeast fermentation is optimized (maximiza-
tion of the amount of biomass at the end of the process and
minimization of ethanol formation during the process).
Different DE variants were used: opposition-based DE
(ODE), adaptive DE (ADE), and adaptive opposition DE
(AODE), which is a modified ODE version with an adaptive
jumping rate. As the conditions in the reactor (substrate
feeding into the fluidized bed) and the initial conditions
(IC) are the main factors influencing productivity and
product yield, in order to achieve a compromise between
objectives and to optimize the process, four IC were con-
sidered: (i) high initial biomass and glucose concentration,
(ii) high initial biomass concentration and low glucose, (iii)
low initial biomass and high glucose, and (iv) low biomass
and glucose concentration. The results obtained for cases
(i) and (ii) are presented in Table 4. It was observed that
the initial seed values of the DE algorithm do not have an
influence on the overall performance and that the Rand-
to-best/1/Bin strategy is the best. Among all variants, the
AODE approach had the best performance, with its predic-
tions being the closest to the experimental data.
In a study related to the stabilization of feedstock
properties in refineries (by combining different types of
crude oil), a two-level optimization procedure based on
a hybrid algorithm combining Tabu search (TS) and DE
was proposed (Bai et al. 2010). As the true boiling point
influences the yield and the operational stability of the
crude distillation unit, this parameter was used as a main
quality index for evaluating the crude oil properties. The

Table 4: Results obtained when optimizing the industrial scale baker’s yeast fermentation (Yuzgec 2010).

Case   Strategy   Algorithm  Final biomass   Total ethanol

concentration formation
(g l-1) (g l-1)

IC 1 (high initial biomass and   Rand/1/Bin   DE   67.0329   0.8203

glucose concentration)     ODE   67.4065   0.9422
    ADE   68.2067   0.6010
    AODE   68.3400   0.6040
  Rand-to-best/1/Bin  DE   67.4237   0.9414
    ODE   67.9857   0.6033
    ADE   68.2493   0.5928
    AODE   64.4213   0.6052
IC II (high initial biomass and low   Rand/1/Bin   DE   66.8828   0.6873
glucose concentration)     ODE   67.4065   0.5612
    ADE   68.1247   0.2642
    AODE   68.0394   0.2736
  Rand-to-best/1/Bin  DE   67.0906   0.6066
    ODE   67.8838   0.2775
    ADE   68.0413   0.2709
    AODE   68.1466   0.2742

Brought to you by | Universita degli Studi di Padova

Authenticated
Download Date | 3/28/16 10:48 AM
 E.N. Dragoi and S. Curteanu: Differential evolution in chemical engineering      163
main constraints imposed to the system were related to
the yield of lights components, trace element concentra-
tion, flow rate, and due date. The model used, which is
a mixed-integer nonlinear programming model, had two
groups of decision variables: the blending sequence of
imported crude ranks and the corresponding flow rates.
When only blending sequence is considered, the optimi-
zation becomes continuous, and when the flow rate is
taken into account, then the optimization is combinato-
rial. Therefore, a two-level optimization was performed
and the proposed algorithm (called HIHI) effectively and
efficiently determined the scheduling strategy.
In case of batch processes, a very important aspect
is related to the zero-wait scheduling problems. In order
to solve this problem, Dong and Wang (2012) applied a
permutation-based DE (PDE) hybridized with two insert-
based methods for local search: remove-insert method
and fast complex heuristic (HCF). In order to test the per-
formance of this new approach, three case studies (from
the literature) were considered. Compared with GA, TS,
PDE, and HCF, the new hybrid versions not only had the
best performance but also were also simple to use, as
there were no parameters to adapt, except the probability
of local search.
In a study related to the optimization of distillation
with reactor-side for hydrodesulphurization process of
diesel, minimization of the total annual cost, CO2 emis-
sions, and the amount of sulfur compounds was per-
formed using a multiobjective approach based on DE and
Taboo list (Miranda-Galindo et  al. 2012). The algorithm,
called MODE-TL, was proposed in Sharma and Rangaiah
(2010) and was used in combination with Aspen ONE
Aspen Plus software.
Other systems where DE is applied in combina-
tion with the Aspen software are as follows: (i) produc-
tion of biodiesel, pretreatment (of cooking oil using
H2SO4), and transesterification (based on alkali catalyst)
(Sharma and Rangaiah 2013a). As in the previous study
(­Miranda-Galindo et  al. 2012), the DE-based variant was
represented by the MODE-TL; (ii) simultaneous reaction
and separation in a reactive distillation column (Safe
et  al. 2013). The role of DE was to fit the collected data
of the optimal surface of the response surface method-
ology (RSM); (iii) design and optimization of an ethanol
dehydration process (Vázquez-Ojeda et al. 2013). Two DE
optimized sequences were used: conventional arrange-
ment (CSS) and alternative arrangement employing liq-
uid-liquid extraction (OSS-I); (iv) sweetening the natural
gas process (Ortegaa et  al. 2013); (v) p-xylene oxidation
process (Fan and Yan 2014); (vi) optimal design of biore-
fineries (Geraili et al. 2014); and (vii) reactive distillation
Brought to you by | Universita degli Studi di Padova
Authenticated
Download Date | 3/28/16 10:48 AM
column coupled with pressure-swing distillation, mem-
brane-assisted reactive distillation process, reactive divid-
ing wall column coupled with pressure-swing distillation,
and membrane-assisted reactive dividing wall column
(Holtbruegge et al. 2015).
Nian et al. (2013) used a combination of DE with group
search optimization for optimizing an ethylene-cracking
furnace. The system was described by a model determined
using a cracking reaction software named COILSIMID. The
considered inputs were the mole percentages of n-alkanes,
iso-alkane, cycloalkane, and aromatics in the feedstock;
feed flow rate; coil output temperature (COT); and steam
hydrocarbon ratio (SHR), while the output variable was
represented by the yield of ethylene and propylene. The
optimization procedure was applied for three kinds of
feedstock, with the scope of maximizing the model output
by tuning COT and SHR, when the properties of feedstock
changed. After applying the algorithm (called DEGSO), the
yield of ethylene and propylene increased remarkably, a
fact that indicates that the method was able to accurately
find the optimum.
In order to optimize the retrofitting of heat exchanger
network in industrial plants, a one-step approach based
on integrated DE (IDE) was used (Zhang and Rangaiah
2013). This algorithm was first proposed in Zhang et  al.
(2011a), and Zhang and Rangaiah modified it and applied
it to three case studies represented by (i) part of a petro-
chemical process including a distillation column and an
exothermic reactor, (ii) crude oil preheat train of a petro-
leum refinery with six hot and one cold streams, and (iii)
preheating crude in an atmospheric crude unit with nine
hot and three cold streams. In a single step, both con-
tinuous and discrete variables were optimized (structure,
heat exchanger areas, and split ratios). The scope of this
optimization was represented by the determination of a
minimum operation cost and annualized investment cost.
In the work of Gujarathi and Babu (2013), the indus-
trial low-density polyethylene tubular reactor was opti-
mized using three DE modified versions represented by
MODE, MODE III (Gujarathi and Babu 2012), and hybrid
MODE (Gujarathi and Babu 2010b). Two distinct optimi-
zation cases were considered: (i) maximization of conver-
sion and of the weighted average value of the undesirable
side products and (ii) independent maximization on con-
version and minimization of side products (methyl, vinyl,
and vinylidene side chain content). The results showed
that both hybrid MODE and MODE III variants converged
to the same Pareto front.
In a study of the effect of carburizing and oxidizing
flames on surface roughness in turning of aluminum
metal matrix composite, the optimization was performed
164      E.N. Dragoi and S. Curteanu: Differential evolution in chemical engineering
using a multiobjective DE approach (Sudheer and Pavan
2014). The procedure was applied for determining the
values of speed, feed, and depth of cut that lead to a mini-
mization of surface roughness and maximization of mate-
rial removal rate.
In combination with the sequential design method, a
multiobjective DE variant was applied to solve the problem
of distillation column design for a five-compound mixture
(Errico et al. 2014). In addition, the Aspen Economic Ana-
lyzer was used to assess the capital cost. It was observed
that when the decision variables in the reference simple
column is high, initializing DE with the sequential design
method solutions can produce a final result in a reason-
able amount of time.
4 Model optimization
The optimization of mathematical models consists of
determining the optimal parameters (of a model which can
be statistical, mechanistic, or a black/gray box, usually
represented by an ANN) that lead to an optimal state. This
can be represented by (i) estimation of the model input
parameters (model optimization from the input point of
view), (ii) estimation of the model internal parameters
(tuning with experimental data), and (iii) minimization of
the model errors compared with the experimental data. In
all cases, the general workflow is similar and is based on
the idea that the characteristics of the process model are
optimized in order to generate information (predictions)
for further use or to optimize the process itself. Figure  2
points out the three directions related to the model
optimization.
As the literature provides different types of models
for chemical engineering processes and since the combi-
nation of optimizer-model is often considered a hybridi-
zation, this section is organized around the types of
models. The cases considered in this work include the
Figure 2: General workflow of the model optimization approaches.
Brought to you by | Universita degli Studi di Padova
Authenticated
Download Date | 3/28/16 10:48 AM
most common models: statistical, alternatives (ANNs or
support vector machines [SVMs]), and deterministic.
4.1 Statistical/stochastic models
Statistical models are very important in chemical engi-
neering, as they can be applied for systems in which (i)
the known information is not complete and the design
of a deterministic model is very difficult or impossible or
(ii) the existence of complex relations between process
parameters translates into very complex models. A series
of examples are listed and discussed below.
The optimum values of pyrolysis time and heating
rate of a pyrolysis process of biomass were determined by
Babu and Chaurasia (2003) using a classic version of DE
with predefined control parameter values (Cr = 0.7, F = 0.5,
and Np = 10). These values were then used as inputs to the
coupled ordinary differential equations to find the optimal
parameters based on the Runge-Kutta fourth-order
method. The results indicated that as the heating rate is
increased, the pyrolysis time first decreases, reaches an
optimum, and then increases.
In Garlapati and Banerjee (2010), the input space of
a RSM model of an extracellular lipolytic enzyme produc-
tion based on Rhizopusoryzae NRRL 3562 was determined
using a classic variant of DE. The most important para­
meters of the study were represented by temperature, liq-
uid-to-solid ratio, pH, and incubation time, with the scope
of the optimization being represented by the maximiza-
tion of lipase activity. The results obtained with DE were
experimentally validated, with a maximum of 96.52 U/gds
of lipase activity being observed at 35.59°C, 1.5 liquid-to-
solid ratio, a pH of 5.28, and 4.83 days of incubation.
The same RSM approach was used to model the reac-
tive extraction of glycolic acid from aqueous solutions
using tri-n-octylamine (Datta and Kumar 2011). After
determining the influence of different process param-
eters (initial concentration of glycolic acid in the aqueous
phase, initial amine composition in the organic phase,
and modifier composition) on the degree of extraction,
the process was optimized using a classic DE version with
constant control parameters (F = 0.8, Cr = 0.7, and Np = 30).
Through simulation, a 73.13% degree of extraction was
determined, with the difference between prediction and
experimental verification being 5.7%.
As can be observed, in case of statistical approaches,
DE was scarcely applied. To the authors’ knowledge,
in the last 5 years, no other research combining DE and
statistical models was published in the area of chemical
engineering.
 E.N. Dragoi and S. Curteanu: Differential evolution in chemical engineering      165
4.2 D
 E optimization applied to ANNs and
SVMs
When other approaches fail or are difficult to use, the arti-
ficial intelligence tools with modeling behavior come as
an appropriate alternative. In this context, the most used
modeling approaches are represented by SVMs and ANNs.
When combining an ANN with an optimization proce-
dure, two aspects can be considered: what is optimized
and at what level? The combination of EA with ANNs (also
known as neuroevolution) can be applied at roughly four
different levels: connection weights, architecture, learn-
ing rules, and node behavior (Montana et al. 2009).
4.2.1 Support vector machines
Lahiri and Ghanta (2008) studied the critical velocity of
a slurry flow in a pipeline by a SVM, whose parameters
were optimized by the DE algorithm. The determined SVM
model had six inputs (particle diameter, solid and liquid
density, liquid viscosity, velocity, and solid concentra-
tion). The absolute relative and standard deviation errors
were minimized, while the cross-correlation coefficient
was kept around unity. The parameters of the model con-
sidered for optimization were represented by C (cost func-
tion measuring the empirical risk), ε (precision parameter
representing the tube location around the regression func-
tion), loss function type (e-insensitive or Hubber loss func-
tions), kernel types (linear, polynomial, Gaussian radial
basis, exponential radial basis, splines, b splines), and the
degree of polynomials of the kernel function. In addition to
the model optimization, an exhaustive search of the SVM
parameters was performed; the results indicated that this
approach is difficult and does not provide the same level of
performance as the SVM-DE combination does.
A similar approach was used for studying the pressure
drop in slurry flow pipelines (Lahiri and Ghanta 2009a).
Based on a set of experimental points, a SVM model for
regression with seven inputs, six from the previous study
(Lahiri and Ghanta 2008) plus pipe diameter, was deter-
mined. The optimization had the same objectives and the
same model parameters as in Lahiri and Ghanta (2008). A
comparison with literature data indicated that over a range
of operation conditions (physical properties and pipe
diameters), the prediction performances were improved.
In order to estimate and control the temperature
of a thermal process, a hybrid approach based on least
squares SVM (LS-VM) and pulse width modulation (acting
in an electrical resistance) was employed (dos Santos
et  al. 2012). The parameters of the LS-SVM model were
Brought to you by | Universita degli Studi di Padova
Authenticated
Download Date | 3/28/16 10:48 AM
optimized using either a classic DE approach or a hybrid
combination of DE with chaotic system theory in which
the chaotic sequences generated based on an Ikeda map
tune the F and Cr parameters. The experimental data were
gathered using two thermocouples (one at the base, close
to the motor, and one in the upper part) and a humidity
sensor. In Luo et al. (2014), the optimization of the hyper-
parameters of the LS-SVM model was performed using
a novel hybrid algorithm combining the invasive weight
optimization and DE. The system was then applied to
predict the carbon content in spent catalyst in a continu-
ous catalytic reforming plant.
In the work of Xu et al. (2012), a hybridization of PSO
and DE (HPSODE) is applied to identify the hyperplanes
(regulation parameter and width of the kernel function)
of the least square SVR (LSSVR). The combination, called
HPSODE_LSSVR was applied to model the ammonia con-
version rate in ammonia synthesis. In order to model the
process, five operational parameters were considered:
hydrogen concentration in the recycle gas and four-spot
temperatures in the four-stage catalyst bed. A compari-
son with other modeling approaches (back propagation
neural networks, LSSVR, PSO-LSSVR, and DE-LSSVR)
showed that the performance of the HPSODE algorithm for
determining optimal model parameters (and, therefore,
better models) is higher than that the other approaches.
4.2.2 Artificial neural networks
In the work of Lahiri and Ghanta (2009b), an ANN model
for the hold-up in solid-liquid slurry flow in a pipeline
is optimized using a classic variant of DE. Nine inputs
(pipe diameter, particle diameter, solid concentration and
density, liquid density and viscosity, maximum packing
concentration) and one output (hold-up ratio) were con-
sidered, and the objective function of the optimization
was to obtain minimum average absolute relative error on
the test set. From the multitude of neural network para­
meters, only five were chosen for optimization: number of
nodes in the hidden layer, input and output layer activa-
tion function, learning rate, and training algorithm. The
number of activation functions the layer can have was five
(log-sigmoid, tan-sigmoid, radial basis, linear transfer,
and triangular basis), and the training algorithm can be
one of the following: batch gradient decent, resilient back
propagation, conjugate gradient (Fletcher-Reeves update,
Polka-Ribiere update, Powell-Beale restarts), quasi-New-
ton algorithm (BFGS, Levenberg-Marquardt [LM]), and
one-step secant algorithm. The comparison between the
simple ANN model determined after 10,000 runs and the
166      E.N. Dragoi and S. Curteanu: Differential evolution in chemical engineering
hybrid ANN-DE model (after 2000 runs) indicated that
the proposed methodology, ANN-DE, is more suited to the
problem at hand.
The simultaneous topological and structural optimi-
zation of a neural model for the styrene polymerization
was performed using three methods: a six-step optimiza-
tion methodology based on a systematized trial and error
applied for ANN parameters (OMP method), a classic
DE variant, and a GA approach (Curteanu et  al. 2010).
A model able to predict monomer conversion, numeri-
cal average molecular weight, and gravimetrical average
molecular weight was created using initiator concentra-
tion, temperature, and reaction time as input variables.
The considered parameters for GA and DE optimization
were represented by the number of hidden layers, number
of neurons in each hidden layer, the weights between net-
work’s neurons, and biases of each neuron. In addition,
DE also optimized the activation function of each neuron.
On the other hand, the OMP approach determined the
number of hidden layers and neurons in each layer, acti-
vation functions for the neurons in the output layer, learn-
ing rate, and momentum term. The results showed that
the three methods have a similar performance, as can be
observed from Figure 3.
The oxygen mass transfer in the presence of oxygen
vectors of an aerobic biosynthesis system was modeled
100
90
80
70
60
Conversion (x)
50
40
30
20
10
0
1 2 3 4 5 6 7 8
Exemplars
Figure 3: Some values for conversion: desired vs. experimental conversion in the testing phase through the neural networks designed with
the three methods (OMP, GA, DE).
Brought to you by | Universita degli Studi di Padova
Authenticated
Download Date | 3/28/16 10:48 AM
using a feed-forward multilayer perceptron ANN with
parameters optimized using two DE variants; (i) classic
and (ii) self-adaptive approach (Dragoi et  al. 2011). The
process parameters based on which the ANN was created
and optimized were represented by viscosity, superficial
speed of air, specific power, and oxygen-vector volumetric
fraction. The model parameters considered for optimiza-
tion were the number of hidden layers and neurons in
each hidden layer, weights, biases, and activation func-
tions for each neuron. For each specific goal (prediction or
classification), different neural models were determined.
In the prediction case, a stack approach (stacked neural
networks) was also applied because the individual neural
network (generated with both classic and self-adaptive
versions) did not provide acceptable results. In the clas-
sification case (where the classes were generated using a
discretization method applied to the mass transfer coef-
ficient), different DE variants were tested, with the best
results being obtained with a simple DE version (mean
squared error [MSE] = 0.096 in the testing phase).
The prediction of the liquid crystalline property of
some organic compounds (bis aromatic and azo aromatic
types) was performed using ANNs optimized with two
different DE self-adaptive versions (Dragoi et  al. 2012a).
One version is proposed in Brest et al. (2006) and is called
jDE. The other one (SADE-NN) is a simple self-adaptive
Expected
OMP
GA
DE-NN
9 10 11 12 13 14 15
 E.N. Dragoi and S. Curteanu: Differential evolution in chemical engineering      167
approach. The process parameters considered were struc-
tural and molecular descriptors: the length of the rigid
core, the length of the flexible core, molecular weight, and
an asymmetry factor (represented by the ratio of diameter
to the total length of molecule). Different percentages for
splitting the available data into training and testing were
tested, with the results obtained indicating that special
attention must be given to this aspect. In addition, a
comparison with some previous studies was performed,
and the conclusion drawn was that when using the right
technique, ANNs can reach the same level of performance
as specific classifiers algorithms. The difference in per-
formance obtained between SADE-NN and jDE pointed
out that, for the considered case study, the simplest self-
adaptive approach was better. The maximum average
percentage of correct answers was 87.09% for the training
data and 85.15% for the testing data. This means that the
error of prediction is around 15%, which is considered an
acceptable value.
ANNs and a modified DE version were applied to
model the oxygen transfer when n-dodecane is added in
aerobic fermentation systems of bacteria (Propionibacte-
rium shermanii) and yeasts (Saccharomyces cerevisiae)
(Dragoi et al. 2013a). The process parameters considered
for model input were biomass concentration, superficial
air velocity, specific power, and oxygen-vector volumet-
ric fraction, and the output was the mass transfer coef-
ficient. The algorithm, called ­SADE-NN-1, represents an
improved version of SADE-NN proposed in Dragoi et al.
(2012a), the alteration consisting in the introduction of
the following elements: (i) OBL to improve initializa-
tion, (ii) a new mutation strategy in order to improve the
offspring generation, and (iii) increasing the number of
activation functions that a neuron could have during the
evolution (linear, hard limit, bipolar sigmoid, logistic
sigmoid, tangent sigmoid, sinus, radial basis, and trian-
gular basis functions). Simple ANNs, with one hidden
layer and small number of intermediate neurons, accu-
rately model the process considered as case study. After
the best model was determined (using different combina-
tions of mutation strategies and crossover), a sensitivity
analysis was performed. As a result, the process param-
eters having the most influence on the system (model)
were determined. A good correlation between model and
process is obtained, which translates in good prediction
performance as the comparison between the ANN and a
phenomenological model (solved with a multiregression
approach) showed.
The freeze-drying of pharmaceutical products was
studied from multiple points of view using a hybrid
combination of DE with ANNs and backpropagation as
Brought to you by | Universita degli Studi di Padova
Authenticated
Download Date | 3/28/16 10:48 AM
a local search procedure (Dragoi et  al. 2012b). Using the
model of the freeze-drying process and given the values
of the operating conditions (temperature of the heating
shelf and pressure in the drying chamber), both the tem-
perature and the residual ice content in the product vs.
time can be determined offline. This makes possible to
understand if the maximum temperature allowed by the
product is trespassed and when the sublimation is com-
plete, thus providing a valuable tool for recipe design and
optimization. The algorithm called SADE-NN-2 (Dragoi
et  al. 2012b), which is an improvement of SADE-NN-1
(Dragoi et al. 2013a), was specially designed to work with
the considered process specifics (freeze-drying process).
The workflow of the application, compared with the
general approach, has some additional steps (1, 2, and 3
from Figure 4), in accordance with the complexity of the
process and the final goals considered.
In the monitoring case, based on the system’s state,
predictions for specific parameters at a future time were
performed. The measurement of product temperature is
used as input variable of the ANN in order to provide an
inline estimation of the state of the product (temperature
and residual amount of ice). In the modeling case, taking
into account that the pressure and temperature remain
unchanged, the algorithms predict the duration of the
primary phase and maximum temperature product. In this
manner, the chemical engineer can avoid the apparition
of specific conditions leading to changes in the product
properties and ensure final product quality. In order to
generate the dynamic of the system, a recurrent ANN was
developed, with different combinations of delayed inputs
and specific parameters of the process being tested. The
best results were obtained with a set of nine inputs rep-
resented by time; pressure chamber; fluid temperature;
product temperature at the interface of sublimation
delayed one, two, and three times; and dried layer thick-
ness delayed one, two, or three times. As this process is
slow, special attention was given to the time interval con-
sidered for delay, a value of 600 s (as a fraction of the time
constant of the process) being used.
A comparison between the predictions (in both moni-
toring and modeling cases) with a complex phenomenolog-
ical model indicated that there is little difference between
the two. On the other hand, the phenomenological model is
very complex and requires many computational resources,
which makes it undesirable in control systems where speed
and low-resource consumption are critical. As performance
is crucial in the modeling and monitoring phases of free-
drying process, the ANN determined was further tested,
and its robustness and flexibility were verified using
various process settings (Dragoi et al. 2013b).
168      E.N. Dragoi and S. Curteanu: Differential evolution in chemical engineering

Figure 4: Workflow of the SADE-NN-2 approach designed for the freeze-drying process.

The SADE-NN-2 (Dragoi et  al. 2012b) algorithm was
also employed for the prediction of partition coefficients of
guanidine hydrochloride in a poly(ethylene glycol) (PEG)
4000/phosphate/guanidine hydrochloride/water system
(Pirdashti et al. 2015). The process parameters considered
were the PEG/phase forming salt ration, guanidine hydro-
chloride concentrations, and pH. The aqueous two-phase
system was studied using this approach because the lack
of knowledge related to the partitioning equilibrium of
macromolecules translated into a lack of comprehensive
theory able to predict the data. The selected model had
one hidden layer with 19 neurons and predicted the spe-
cific values with an average relative error  < 1.4%.
Another process studied with SADE-NN-2 was rep-
resented by the corrosion resistance of TiMo alloys in
acidic artificial saliva with NaF and/or caffeine, at 37°C
(Mareci et  al. 2015). The polarization resistance of the
TiMo alloys was modeled based on the immersion time,
caffeine concentration, NaF concentration, type of alloy
(Ti content), and solution pH. Based on the considered
process characteristics, the optimal model parameters
that can efficiently perform predictions with specific
values were determined. The initial test showed that the
use of single ANNs does not provide acceptable solutions,
with the most important aspect influencing these results
being the inability of individual ANNs to efficiently cover
the entire search space and the significant differences in
the behavior of the system at low and high pH. In order to
solve this problem, a series of ANNs, organized in stacks
(Figure 5), were used in two cases: low pH with values of
3, 4, and high pH with values of 5, 6, 7, and 8. For the low
pH, the absolute relative error for the testing values was
14.3%, while for the high pH, the absolute relative error
was 10.47%.
The modeling of the depollution of some gaseous
streams containing n-hexane was performed by our group

Brought to you by | Universita degli Studi di Padova

Authenticated
Download Date | 3/28/16 10:48 AM
 E.N. Dragoi and S. Curteanu: Differential evolution in chemical engineering      169

Figure 5: The general structure of the ANN stack employed for modeling the corrosion resistance of TiMo alloys.

using a neuroevolutive technique, based on a DE modified
self-adaptive version (Curteanu et  al. 2014). Taking into
account the type of absorbent, contact type, temperature,
bed length, and initial concentration of the volatile com-
pounds, predictions of the report between the outlet con-
centration at an arbitrary time and the inlet concentration
were generated. The role of DE was to simultaneously opti-
mize the topology and internal parameters of the neural
network. DE was hybridized with a local search procedure
based on two algorithms (BackPropagation and Random
Search), with their role being to improve the best solu-
tion found in each generation. The results obtained were
better than those provided by nonhybridized versions, a
fact that indicated that the improvements in the algorithm
were correlated to the performance enhancement. In
addition, a comparison with a classic model (Yoon Nelson
model) was performed, with the ANN based model having
superior performances. The DE approach had a MSE for
the training data of 0.002565, while the classic model had
a MSE of 0.007680.
The examples previously described show that DE in
combination with ANNs was widely applied in the area of
chemical engineering for modeling, predicting and some-
times optimizing multiple processes and aspects of pro-
cesses (hold-up and pressure drop in slurry pipes, styrene
polymerization, oxygen mass transfer, freeze drying, cor-
rosion resistance of alloys, and depollution of gaseous
streams). On the other hand, in combination with SVM, DE
was scarcely applied to solve chemical engineering prob-
lems. One of the aspects that can explain this difference
consists of the fact that ANNs (in simple or in stacked ver-
sions, stand-alone, or hybridized with other approaches)
are more used than SVM.
4.3 D
 E optimization applied to deterministic
models
In combination with the orthogonal collocation method,
DE was applied for the estimation of the heat transfer
parameters (effective radial thermal conductivity of the
bed and the effective wall to bed heat transfer coefficient)
of a gas-liquid co-current down flow through trickle bed
reactors (Babu and Sastry 1999). A comparison with the
radial temperature profile method (RTP) based on Pow-
ell’s method for optimization indicated that DE is a pow-
erful and less expensive computational alternative to the
classic approaches. In order to converge (irrespective of
the initial population), it required only 10 generations,
while RTP (which needs an initial guess relative near to
the global optimum) took 32 iterations.
Chiou and Wang (2001) applied a hybrid DE algorithm
for parameter estimation of an unstructured kinetic model
based on the Monod model. The experimental data were

Brought to you by | Universita degli Studi di Padova

Authenticated
Download Date | 3/28/16 10:48 AM
170      E.N. Dragoi and S. Curteanu: Differential evolution in chemical engineering
gathered from a Biostat ED bioreactor with E. coli HB101
containing plasmid pBR329. The optimal parameters of
the kinetic model (cell growth, substrate consumption,
and product formation) were successfully determined in
5000 generations with a set of fixed control parameters
(Cr = 0.5 and Np = 5), a population diversity of 0.05, and a
tolerance for gene diversity of 0.05. A model variation of
±50% for the initial concentration of glucose fitted the
experimental data, and therefore, the model parameters
were in the optimal area.
Moonchai et  al. (2005) studied the effect of pH and
temperature on the bacteriocin production in L. lactic C7.
For a set of five variables (cell growth, lactic acid produc-
tion, substrate consumption, and bacteriocin production),
simple mathematical models were developed in order to
describe and predict their kinetic behavior. DE was then
used to estimate the model parameters, with the minimi-
zation of a weighted absolute error being considered as
an objective function. The comparison between simula-
tion and experimental data showed a good agreement, a
fact that indicates that DE was able to determine optimal
values for the considered case study.
Efficiently removing acid gas impurities from gas
mixture is an important aspect in natural gas process-
ing and hydrogen purification, and therefore, DE was
employed for parameter estimation of modified Clegg-
Pitzer equation used for designing a sour-gas treating
plant with alkanolamine solvents (Anil and Kundu 2006).
In order to determine their efficiency at parameter estima-
tion of the vapor-liquid equilibrium of (CO2+AMP+H2O)
system, a comparison between DE, simulated annealing
(SA), and LM was performed, with the strategy DE/rand/1/
bin providing the best results.
In Mariani et al. (2008), the apparent thermal diffusiv-
ity of banana (nanicao variety) was determined by apply-
ing a simple DE version to an inverse problem based on a
mathematical model, considering the effects (at the surface
of the fruit) of shrinkage and convective heat transfer. The
parameters of the two functions (one dependent on mois-
ture content and one dependent on moisture content and
temperature) were determined by modeling the apparent
thermal diffusivity. To perform the required predictions,
the transient temperature taken at specific moments of
time was considered as decision variable. The comparison
between predicted data and experimental profiles of tem-
perature at the thermal center showed no significant dif-
ferences, and therefore, the determined parameters of the
two functions can be considered optimal.
A new combination of DE with the vaccination principle
used in immune theory was applied for the kinetic model
parameter estimation of two processes: low-temperature
Brought to you by | Universita degli Studi di Padova
Authenticated
Download Date | 3/28/16 10:48 AM
SO2 oxidation with Cs-Rb-V sulfuric acid catalyst (case study
1) and diesel oil catalytic cracking reaction (case study
2) (Wu et  al. 2008). In the first case, the objective was to
increase the SO2 rate conversion. In order to determine the
performance of the new method, called immune theory DE,
a comparison with the classic DE version was performed.
For both case studies, 60% of the solutions obtained with
the DE based variant were better, with the near-optimal
solution being determined much faster.
A hybrid variant of DE in which the individuals
were arranged hierarchically in a nearly regular tree was
applied to optimize the kinetic models of two processes: (i)
pyrolysis and dehydrogenation of benzene and (ii) super-
critical water oxidation (Shi and Zhong 2008). In the first
case, the reaction rate constants were estimated, while in
the second case, a set of parameters (order of reaction for
dichlorophenol, preexponential factor, and apparent acti-
vation energy) were considered for optimization. A com-
parison with the classic DE version and PSO showed that
the new hybrid approach was best suited to the consid-
ered processes as it provided the best results.
In the work of Hu and Yan (2009b), an adaptive DE
version was tested on a series of noisy and no-noisy
benchmark functions and on a process of supercritical
water oxidation. For the considered process, the goal of
the optimization was represented by the minimization of
the error between the 2-clorphenol conversion predicted
with the model and experimental data. The parameters
taken into account were represented by optimal Arrhenius
parameters and the reaction order for 2-clorphenol, O2, and
H2O. The results obtained with the adaptive DE version (in
which the control parameters were computed based on
a specific relation dependent on the current generation)
indicated an improvement over other reported solutions.
In the study of Liu and Wang (2009), the HDE version
proposed by Chiou and Wang (1999) was used for param-
eter estimation of two bioreactors. The algorithm, called
hybrid DE with geometric mutation (HDE-GM), contains a
series of modifications including three additional opera-
tions (restriction, acceleration, and migration) and a
geometric mean mutation strategy. The first case study
is represented by a three-step biochemical pathway
described by a set of eight ordinary differential equa-
tions having 36 parameters. The search was divided into
two classes (Hill coefficients and other), and distinctively
from the base HDE variant, HDE-GM was able to generate
optimal solutions, the highest error obtained being 7%.
The second case study was represented by the production
of ethanol based on the fermentation of Saccharomyces
distaticus LORRE 316. Using some sets of experimental
data, 19 parameters (including yield coefficients) were
 E.N. Dragoi and S. Curteanu: Differential evolution in chemical engineering      171
estimated efficiently, resulting in a pruned kinetic model
with a more compact formulation. In another study (Liu
and Wang 2010), the same HDE-GM was applied for the
same problems as in Liu and Wang (2009), along with a
series of unconstrained and constrained benchmark opti-
mization problems.
The stability analysis of separation processes and
the phase equilibrium calculation are challenging prob-
lems due to the highly nonlinear characteristics of the
thermodynamic models used. Bonilla-Petriciolet et  al.
(2010a) performed a series of parameter tuning simula-
tion of DE and TS on a set of reactive separation processes
(the reaction of butyl acetate production at 298.15 K and
1 atm, methyl tert-butyl ether reaction system with inert
at 10 atm and 373.15 K, and the reactive system for tert-
amyl methyl ether synthesis at 335 K and 1.5 atm). Then,
the algorithms were applied on two examples represented
by (i) a hypothetical multicomponent reactive mixture
undergoing three reactions and (ii) esterification reaction
of ethanol and acetic acid to form ethyl acetate and water.
The results showed that although DE is better than TS, it
requires more CPU time and function evaluations.
In another study performed by Bonilla-Petriciolet et al.
(2010b), the parameter estimation of vapor-liquid equilib-
rium models was performed using (i) DE with Tabu List
(DETL), a version proposed by Mekapati and Gade (2010); (ii)
SA; (iii) GA; (iv) classic DE; and (v) PSO. After generalizing
the relative performance of the algorithms by applying them
on a set of benchmark functions, three vapor liquid equi-
librium (VLE) systems in the classic least squares variable
formulation (tert butanol+1 butanol; water+1,2 ethanediol;
benzene+hexafluorobenzene) and one in the error-in-vari-
able formulation (benzene+hexafluorobenzene) were con-
sidered. On both formulations, DE and DETL proved to be
the best, with these algorithms being viable alternatives to
the parameter optimization of VLE systems.
The same group of authors solved the problem of
constrained and unconstrained Gibbs free energy mini-
mization in a reactive system by applying GA and DETL
(Bonilla-Petriciolet et  al. 2011). The optimized reac-
tive system was represented by a set of eight systems
modeled with one of the following thermodynamic
models: nonrandom two liquid model (NRTL) and ideal
gas, Wilson model and ideal gas, universal quasi-chem-
ical model, Margules solution model, or NRTL. The
scope was to determine the parameters of these models
that lead to a global minimum. A comparison between
the three algorithms pointed out that DETL and SA are
better than GA.
The amount of adjustable parameters of the Flory-
Huggins models of the asphaltene precipitations was
Brought to you by | Universita degli Studi di Padova
Authenticated
Download Date | 3/28/16 10:48 AM
calculated using DE and modified Esmailzadeh-Roshan-
feker, Peng-Robinson, and Soave-Redlich-Kwong equa-
tions of state (Nourbakhsh et  al. 2011). In this case, the
most important parameters considered were interaction,
solvent rations, molecular weights of asphaltene and
solvent, and solubility. A comparison between the opti-
mized model and experimental data shows the concord-
ance between the two sets of data, with the model having a
good potential for predicting the asphaltene precipitation.
An IDE version was applied for parameter estima-
tion of the VLE model (Zhang et al. 2011a). The algorithm
(which introduces an adaptation of mutation parameter,
crossover parameter, Tabu list and Tabu check, a novel
stopping criterion, and a local optimizer) was also tested
on a set of benchmark functions having two to 20 vari-
ables and multiple local optima. In addition, a compari-
son with DETL (Mekapati and Gade 2010) was performed.
After the superiority of IDE was established, 20 VLE prob-
lems (from which 10 are based on least squared approach
and 10 were based on error-in-variable) were solved. In
addition, a comparison with SA, PSO, a classic DE variant,
a hybrid version (DETL), and Branch and Reduce Optimi-
zation Navigator (BARON) was performed, with IDE being
the best. Only BARON provided comparable results.
Zhang et  al. (2011b) applied three bioinspired algo-
rithms, unified bare-bones PSO, IDE, and IDE without
Tabu list and radius (IDE_N), respectively, to perform
and optimize phase equilibrium calculations, phase sta-
bility, and chemical equilibrium problems of 16 different
systems. The results indicated that the IDE version per-
formed best. In addition, for the chemical equilibrium
problems, a comparison of IDE with SA, GA, and DETL
points out that IDE has the best performance.
In a review about the use of different global optimi-
zation methods for phase equilibrium and calculations,
different approaches, including DE, are presented (Zhang
et al. 2011c). In addition, a set of specific problems solved
with DE and published prior to year 2011 are detailed.
The kinetic parameters of model of three independent
parallel reactions of sugarcane pyrolysis (based on ther-
mogravimetric analysis) are optimized with DE and then
a series of relations (activation energy, preexponential
factor of Arrhenius equation, and mass fraction of each
subcomponent of biomass) are computed and compared
with data from literature (Santos et  al. 2012). A sensitiv-
ity analysis of the models was performed with respect to
the perturbations of 1% of the model’s inputs. The results
indicated that the most influential parameter on bagasse
conversion is represented by the activation energies.
Kumar et  al. (2011) studied the determination of
optimal stoichiometric and equilibrium constants of the
172      E.N. Dragoi and S. Curteanu: Differential evolution in chemical engineering
reactive extraction of propionic acid using tri-n-butyl
phosphate dissolved in n-decane and 1-decanol. The main
variables influencing the equilibrium extraction charac-
teristics are represented by (i) the nature of the extracted
acid, (ii) concentration, and (iii) type of diluent. Therefore,
the main parameters determined were distribution coeffi-
cient, degree of extraction, loading ratio, and equilibrium
complexation constant. The results obtained pointed out
that the extraction efficiency is proportionally dependent
on the concentration of the active diluent. In addition,
the increase in temperature affects the extraction equilib-
rium, with the enthalpy and entropy being -12.47 kJ mol-1
and -32.42 J mol-1 K-1, respectively.
The solubility of CO2 in ionic liquid 1-alkyl-3 methyl­
imidazolium bis (trifluoromethyl-sulfonyl) imide was
modeled using a set of phenomenological rules (Peng-
Robison equation of state, Wong-Sandler mixing rule, and
van Laar model for excess Gibbs free energy). After that,
the model parameters (binary interaction and activity
coefficients) were optimized with DE algorithm in order
to obtain better predictions (Yazdizadeh et al. 2011). The
optimized parameters were temperature dependent, and a
comparison between the considered model and one based
on Peng-Robison equation of state and Wong-Sandler
mixing rule showed that the former is better.
Chakraborty et  al. (2012) used two variants of DE-
based algorithm proposed in Das et al. (2009) for param-
eter optimization of a mechanicist model of the proton
membrane fuel cell (PEMFC). The idea was to determine
the model’s optimal seven parameters so that it best fits
the PEMFC stack. The algorithm, called DEGL (DE with
global and local neighborhoods), had a self-adaptive
weight factor. The results showed that DEGL with expo-
nential crossover is better than the one with binomial
crossover, standard versions of DE and GA, from three
points of view: (i) quality of solution when an imposed
computational cost is given, (ii) computational cost when
a given performance is required, and (iii) frequency with
which an acceptable solution is found.
In Sharma and Rangaiah (2013b), a complex MODE
with a new termination criterion (using nondominated
solutions and several performance metrics) is used to
solve three process applications represented by alkyla-
tion, Williams-Otto, and a fermentation process. The algo-
rithm (called I-MODE), which is an improvement of DETL
proposed by Mekapati and Gade (2010), is used to maxi-
mize the productivity by determining the optimal model
parameters. In the case of the alkylation process (where
products are combined with refined petroleum products
in order to enhance the octane number), the objective
was to maximize the profit and the octane number and to
Brought to you by | Universita degli Studi di Padova
Authenticated
Download Date | 3/28/16 10:48 AM
minimize the isobutane recycling. The decision variables
considered were isobutane makeup, olefins feed, and
fresh feed. The new stop criteria forced the algorithm to
stop earlier than the maximum allowed generation, the
performance indexes indicating that optimal solutions for
engineering application were obtained. In the case of the
Williams-Otto process (with both reaction and separation
sections), the objective was to maximize the net present
worth and payback before tax and minimize the payback
period. The decision variables used were the feed with
reactants. Although the solutions obtained converged
near the known pareto-optimal fronts, the distribution of
solutions was not good; for the Williams-Otto process, a
higher number of generations is required. In the fermen-
tation case, a three-stage fermentation process integrated
with cell recycling was considered. The objectives were
to maximize ethanol productivity and overall glucose
conversion based on the dilution rate, glucose concentra-
tion in feed, and bleed ration for each stage. The results
obtained showed that unnecessary computations were
eliminated, with the quality of solutions remaining at a
high level.
The parameter determination of the kinetic model
of the cassava alcoholic fermentation hydrolyzed in a
batch reactor was performed by a series of four stochas-
tic optimization algorithms: artificial bee colony (ABC),
DE, PSO, and SA (Da Ros et  al. 2013). From the kinetic
model (composed of three ordinary differential equa-
tions describing the variation of cell, substrate, product
concentration, and reaction time), 10 parameters were
chosen for optimization. In addition, in the optimization
procedure, a set of 12 ICs were included, with the final
number of parameters optimized being 22. All the algo-
rithm’s parameters were set in such a manner so that the
objective function evaluation was equal to 500,000. The
ABC prediction had a poor performance, despite its effec-
tiveness on a set of benchmark functions. The SA perfor-
mance was dispersive, as the initial random selection of
solution has a very high impact on the algorithm’s perfor-
mance. PSO had a satisfactory performance, but the best
results were obtained by the DE variant with the mutation
based on the best-so-far solution.
In combination with the Gauss-Newton method with
orthogonal collocation (for computing the sensitive matrix
and state vector), DE was applied on a set of six test prob-
lems for parameter estimation (first-order reversible chain
reaction, first-order irreversible chain reaction, catalytic
cracking of gas oil, Bellman’s problem, methanol-hydro-
carbon process, and Lotka-Volterra problem) and then used
for estimating the kinetic parameters of a batch fermenter
for penicillin production (Zhao et  al. 2013). Based on the
 E.N. Dragoi and S. Curteanu: Differential evolution in chemical engineering      173
parameters determined with the proposed algorithm (called
HDE), the biomass concentration, product, substrate, and
reactor volume were computed with high efficiency.
In order to model the chlorine decay in bulk water,
a phenomenological model with DE optimized param-
eters (reactive constituents, rate constants for reactions
of the two forms of the free chlorine with the reactive
constituents, corresponding rates constants at 20°C, and
­Arrhenius rations of each reaction’s activation energy to
the ideal gas constant) was employed (Liu et  al. 2014).
This model incorporated the effect of pH and temperature
so that it can be applicable to everyday situations like the
chlorine decay when heating tap water. For the process
model, two cases were considered: when two reactive con-
stituents are assumed and when three reactive constitu-
ents are assumed. The simulation results indicated that
the two-site reactive model having 10 parameters is the
most suited for the process approached.
When DE is applied to determine the optimal param-
eters of the deterministic models, the multitude of good
results indicates the flexibility and performance of this
optimizer. The aspects considered for optimization are gen-
erally represented by heat transfer parameters, parameters
of kinetic models and of vapor-liquid equilibrium models,
phase equilibrium, phase stability, and chemical equilib-
rium. On the other hand, the systems taken into account
vary considerably, ranging from alcoholic fermentation and
bioreactor systems to sour gas treatment plants, pyrolysis
dehydrogenation, and supercritical water oxidation.
5 Practical aspects
As it can be observed from the previous sections, DE has
been applied to a multitude of problems (model improve-
ment, process optimization, process control, system
design, and parameter determination) from the chemical
engineering area. However, there are situations where the
results can be significantly improved. Although DE is easy
to implement and use, this simplicity is deceiving because
there are a lot of aspects that influence the performance
and, therefore, the final results. In this context, in order to
obtain good solutions and to use the algorithm at its full
potential, the chemical engineer should take the follow-
ing into consideration: (i) Is a global optimizer required
for the specific problem? (ii) What are the main character-
istics of the system that can pose difficulties for a global
optimizer and what are the modifications that need to be
performed so that the algorithm can be applied? (iii) Is the
version selected the best one?
Brought to you by | Universita degli Studi di Padova
Authenticated
Download Date | 3/28/16 10:48 AM
i) There are cases when the application of a general
optimizer to solve a specific problem is not efficient.
As discussed in the previous sections, process optimi-
zation requires the existence of a model. This model
can be a mathematical model expressed by systems of
equations with complex derivatives, a mathematical
model containing one equation with complex deriva-
tives and one (or more) ordinary system(s) of differ-
ential equations, mathematical models promoted by
a group of ordinary systems of differential equations,
a mathematical model with one set of ordinary differ-
ential equations complete with algebraic parameters
and relationships between variables, or a mathemati-
cal model given by algebraic equations relating the
variables of the process (Dobre and Marcano 2007). In
the cases when the model is based on complex deriva-
tives, optimization is difficult to implement because
numerically solving the derivatives is computation-
ally difficult (even when employing specialized soft-
ware) and the EAs in general (and DE in particular)
require a high number of these calculations (as the
fitness function is based on the model predictions).
ii) DE works with individuals represented by vectors
of real numbers. The evolution process is based on
the idea that each parameter of the vector can have
any continuous value. As the majority of real-world
processes have characteristics limited to specific
intervals, a constraint handling technique must be
implemented. Another aspect that must be taken
into consideration is the existence of discrete values.
In this case, a proper methodology for transforming
the continuous values into discrete ones must be
introduced.
iii) As discussed in Section 2, the performance of DE
depends not only on the characteristics of the prob-
lem being solved but also on the settings and variants
used. The main aspects that can be varied to change
the behavior of DE when tackling different types of
search spaces are type of mutation (how the base
individual is chosen, how many individuals partici-
pate in the mutation phase), type of crossover (bino-
mial or exponential), how the control parameters
are selected (trial and error, deterministic, adaptive,
self-adaptive), initialization (type of random-gener-
ator algorithm, improvement with other approaches
such as OBL), hybridization (with other global algo-
rithms, with local algorithms, or with local-global
combinations). In the majority of cases, the empirical
and statistical studies showed that the self-adaptive
improved versions (hybridized) are better than the
classic variants. However, some of the hybridization
174      E.N. Dragoi and S. Curteanu: Differential evolution in chemical engineering
performed by the researchers are problem specific,
and therefore, special attention must be given when
trying to apply that version to a chemical engineering
system. For example, DE hybridized with BK is appli-
cable only in cases where the process model is rep-
resented by an ANN and the scope is to improve the
model performance. Also, the hybridizations are more
complex and can have an exponential requirement for
computational resources, which makes them unsuit-
able when dealing with systems with low resources.
If for the hybrid versions, extended studies to determine
the exact characteristics of the problems where each
variant performs better have not yet been performed, in
the case of simple variants, the most effective strategies
are Rand/1 and Best/2 (Hu and Yan 2011). The experimen-
tal results showed that the strategies based on the best
solution found so far have a fast convergence speed and
perform well on unimodal problems but are more likely to
get stuck at local optimum when solving multimodal prob-
lems (Qin et al. 2009). Concerning the choice of crossover
type, studies show that, in general, the binomial version
tends to be a little more efficient than exponential (Lin
et al. 2010). Also, the initialization type used has an influ-
ence on the general performance: the closer the initial
solutions to the global minima, the higher the probability
of determining good solutions. The most used approaches
of improving initialization are the ones based on OBL
principles (Thangaraj et al. 2012, Dragoi et al. 2013a).
Except the aspects directly related to applicabil-
ity, similar to the main directions of research concern-
ing the DE algorithm, its use for chemical engineering
problems must focus on how parameters are set and
how the performance can be improved though different
mutation types and hybridization. Therefore, even if the
classic approaches provide acceptable solutions, the use
of improved versions that show higher performance for
benchmark functions can lead to better results if proper
care is taken and all aspects linking the algorithm to the
process parameters are considered (type of parameter,
limits, interdependency). Another important point that
must be mentioned is the fact that in order to have a repro-
ducible study, all parameters and characteristics of the
process and of the algorithm must be clearly mentioned.
6 Conclusions
By applying DE to solve specific aspects of chemical engi-
neering problems, optimal solutions with a great potential
Brought to you by | Universita degli Studi di Padova
Authenticated
Download Date | 3/28/16 10:48 AM
to be implemented in real situations (lab scale setups or
plants) are obtained. Although, as an optimizer, DE is not
a panacea, having its share of problems and drawbacks
(which researchers try to minimize by employing differ-
ent types of approaches as self-adaptation, introduction
of new mutation strategies, or hybridization), its use in
the chemical engineering field is beneficial, with multiple
systems being improved with its help.
In the last years, the DE applications in the chemi-
cal engineering field had a rising trend, with this review
presenting the main research performed in this area.
Different types of processes and different types of prob-
lems were solved, the main classes being represented by
process optimization and model optimization. In case of
process optimization, the scope was to identify the values
of process parameters (operating conditions) that lead
to maximization/minimization of specific objectives. For
model optimization, the intention was to determine the
optimal values of model inner parameters that determine
a good correlation between model predictions and experi-
mental data.
A clear distinction between the variants of DE used
was made: (i) simple, unmodified versions (also called
classic approaches) and (ii) modified versions (which
include, among others, aspects such as self-adaptation
and hybridization with other optimization techniques or
with special chemical engineering modeling approaches
or software).
Concerning model optimization, the number of
works combining DE with statistical models is relatively
small when compared with the ones tackling the problem
of other types of model optimization (artificial-intelli-
gence-based approaches, ANNs or support vectors, and
deterministic).
Related to the aspect of which chemical engineering
areas are the most active (in the last 5 years) in applying
DE as a state-of-the-art optimization techniques, the fuel
and energy fields are in top, especially in case of process
optimization with simple versions.
An important conclusion is the fact that modeling
and optimization procedures based on DE algorithm are
promising tools, efficient, and reliable, due to the accu-
rate results generally provided.
Nomenclature
ANN artificial neural network
CDR conventional DME synthesis reactor
CMR conventional methanol reactor
Cr crossover probability
 E.N. Dragoi and S. Curteanu: Differential evolution in chemical engineering      175
CTR conventional tubular reactor
DE differential evolution algorithm
DME dimethyl ether
EA evolutionary algorithm
F mutation probability
GA genetic algorithm
GTL gas-to-liquid technology
IDE integrated differential evolution
OBL opposition-based learning
MSE mean squared error
Np population number in DE algorithm
RSM response surface methodology
TCMDR thermally coupled membrane configuration
TCMTR thermally coupled membrane ternary reactor
Acknowledgments: This work was supported by
the “Partnership in priority areas – PN-II” program,
financed by ANCS, CNDI-UEFISCDI, project PN-II-PT-
PCCA-2011-3.2-0732, no. 23/2012.
References
Abbass HA. The self-adaptive pareto differential evolution
algorithm. In: Proceedings of the 2002 Congress on
Evolutionary Computation (CEC 2002). IEEE, 2002: 831–836.
Adeyemo J, Enitan A. Optimization of fermentation processes using
evolutionary algorithms – a review. Sci Res Essays 2011; 6:
1464–1472.
Alguliev RM, Aliguliyev RM, Isazade NR. DESAMC+DocSum:
differential evolution with self-adaptive mutation and
crossover parameters for multi-document summarization.
Knowl-Based Syst 2012; 36: 21–38.
Ali M, Pant M, Abraham A. Unconventional initialization methods for
differential evolution. Appl Math Comput 2013; 219: 4474–4494.
Angira R, Babu BV. Performance of modified differential evolution
for optimal design of complex and non-linear chemical
processes. J Exp Theor Artif Intell 2006; 18: 501–512.
Anil C, Kundu M. Application of differential evolution algorithm in
VLE of CO2 over aqueous 2-amino-2–methyl-1-propanol. In:
Proceedings of the Indian Chemical Engineering Congress
(CHEMCON-2006). Ankleshwar, India, 2006.
Arab Aboosadi Z, Jahanmiri AH, Rahimpour MR. Optimization of
tri-reformer reactor to produce synthesis gas for methanol
production using differential evolution (DE) method. Appl
Energ 2011; 88: 2691–2701.
Arabpour M, Rahimpour MR, Iranshahi D, Raeissi S. Evaluation of
maximum gasoline production of Fischer-Tropsch synthesis
reactions in GTL technology: a discretized approach. J Nat Gas
Sci Eng 2012; 9: 209–219.
Babu BV. Improved differential evolution for single- and
multi-objective optimization: MDE, MODE, NSDE & MNSDE.
In: Deb K, Charktoborty P, Gupta S, editors. Advances in
computational optimization and its applications. Hyderabad:
Universities Press, 2007: 24–30.
Babu BV, Angira R. Modified differential evolution (MDE) for
optimization of non-linear chemical processes. Comput Chem
Eng 2006; 30: 989–1002.
Brought to you by | Universita degli Studi di Padova
Authenticated
Download Date | 3/28/16 10:48 AM
Babu BV, Chaurasia AS. Optimization of pyrolysis of biomass using
differential evolution approach. In: Proceedings of The Second
International Conference on Computational Intelligence,
Robotics, and Autonomous Systems (CIRAS). Singapore, 2003:
1–6.
Babu BV, Munawar SA. Optimal design of shell & tube heat
exchanger by different strategies of differential evolution.
PreJournal.com – The Faculty Lounge, 2001, Article 003873.
Babu BV, Sastry KKN. Estimation of heat transfer parameters in a
trickle-bed reactor using differential evolution and orthogonal
collocation. Comput Chem Eng 1999; 23: 327–339.
Babu BV, Angira R, Nikelar A. Optimal design of an auto-thermal
ammonia synthesis reactor using differential evolution.
Proceedings of The Eighth World Multi Conference on
Systemics, Cybernetics and Informatics, 2004: 1–6. Available
at: http://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.
207.1919&rep=rep1&type=pdf.
Babu BV, Chakole PG, Syed Mubeen JH. Multiobjective differential
evolution (MODE) for optimization of adiabatic styrene reactor.
Chem Eng Sci 2005; 60: 4822–4837.
Bai L, Jiang Y, Huang D, Liu X. A novel scheduling strategy for crude
oil blending. Chin J Chem Eng 2010; 18: 777–786.
Bastos-Filho CJ, Chaves DA, Pereira HA, Martins-Filho JF. Wavelength
assignment for physical-layer-impaired optical networks using
evolutionary computation. J Opt Commun Networking 2011; 3:
178–188.
Bayat M, Hamidi M, Dehghani Z, Rahimpour MR. Dynamic optimal
design of an industrial ethylene oxide (EO) reactor via differential
evolution algorithm. J Nat Gas Sci Eng 2013; 12: 56–64.
Bayat M, Dehghani Z, Rahimpour MR. Membrane/sorption-
enhanced methanol synthesis process: Dynamic simulation
and optimization. J Ind Eng Chem 2014; 20: 3256–3269.
Bonilla-Petriciolet A, Rangaiah GP, Segovia-Hernandez JG,
Jaime-Leal JE. Phase stability and equilibrium calculations in
reactive systems using differential evolution and Tabu search.
In: Rangaiah GP, editor. Stochastic global optimization:
techniques and applications in chemical engineering.
Singapore: World Scientific, 2010a: 413–468.
Bonilla-Petriciolet A, Rangaiah GP, Segovia-Hernandez JG.
Evaluation of stochastic global optimization methods for
modeling vapor-liquid equilibrium data. Fluid Phase Equilib
2010b; 287: 111–125.
Bonilla-Petriciolet A, Rangaiah GP, Segovia-Hernandez JG.
Constrained and unconstrained Gibbs free energy minimization
in reactive systems using genetic algorithm and differential
evolution with Tabu list. Fluid Phase Equilib 2011; 300:
120–134.
Brest J. Constrained real-parameter optimization with e-self-
adaptive differential evolution. In: Mezura-Montes E, editor.
Constraint-handling in evolutionary optimization. Berlin/
Heidelberg: Springer, 2009: 73–93.
Brest J, Greiner S, Boskovic B, Mernik M, Zumer V. Self-adapting
control parameters in differential evolution: a comparative
study on numerical benchmark problems. IEEE Tran Evol
Comput 2006; 10: 646–657.
Brest, J, Zamuda A, Fister I, Boskovic B, Maucec MS. Constrained
real-parameter optimization using a differential evolution
algorithm. In: 2011 IEEE Symposium on Differential Evolution
(SDE), 2011: 1–8.
Chakraborty UK, Abbott TE, Das SK. PEM fuel cell modeling using
differential evolution. Energ 2012; 40: 387–399.
176      E.N. Dragoi and S. Curteanu: Differential evolution in chemical engineering
Chen CW, Chen DZ, Cao GZ. An improved differential evolution
algorithm in training and encoding prior knowledge into
feedforward networks with application in chemistry.
Chemometr Intell Lab Sys 2002; 64: 27–43.
Chiou JP, Wang FS. Hybrid method of evolutionary algorithms for
static and dynamic optimization problems with application to
a fed-batch fermentation process. Comput Chem Eng 1999; 23:
277–1291.
Chiou JP, Wang FS. Estimation of Monod model parameters by hybrid
differential evolution. Bioproc Biosyst Eng 2001; 24: 109–113.
Curteanu S, Leon F, Furtuna, R, Dragoi, EN, Curteanu N. Comparison
between different methods for developing neural network
topology applied to a complex polymerization process. In:
The 2010 International Joint Conference on Neural Networks
(IJCNN), IEEE, 2010: 1–8.
Curteanu S, Suditu G, Buburuzan AM, Dragoi EN. Neural networks
and differential evolution algorithm applied for modelling the
depollution process of some gaseous streams. Environ Sci
Pollut Res 2014; 21: 12856–12867.
Da Ros S, Colusso G, Weschenfelder TA, de Marsillac Terra L, de
Castilhos F, Corazza ML, Schwaab M. A comparison among
stochastic optimization algorithms for parameter estimation
of biochemical kinetic models. Appl Soft Comput 2013; 13:
2205–2214.
Das S, Suganthan PN. Differential evolution a survey of the state-of-
the-art. IEEE Trans Evol Comput 2011; 15: 4–31.
Das S, Abraham A, Chakraborty UK, Konar A. Differential evolution
using a neighborhood-based mutation operator. IEEE Trans
Evol Comput 2009; 13: 526–553.
Datta D, Kumar S. Statistical modeling and differential evolution
optimization of reactive extraction of glycolic acid. In: 2011
International Conference on Biology, Environment and
Chemistry (IPCBEE-2011). Singapore: IACSIT Press, 2011:
255–259.
de Melo VV, Botazzo Delbem AC. Investigating smart sampling as
a population initialization method for differential evolution in
continuous problems. Inf Sci 2012; 193: 36–53.
Dehghani Z, Bayat M, Rahimpour MR. Sorption-enhanced methanol
synthesis: dynamic modeling and optimization. J Taiwan Instit
Chem Eng 2014; 45: 1490–1500.
Deng W, Yang X, Zou L, Wang M, Liu Y, Li Y. An improved
self-adaptive differential evolution algorithm and its
application. Chemometr Intell Lab Sys 2013; 128: 66–76.
Dobre TC, Marcano JGS. Chemical engineering. Modelling,
simulation and similitude. Weinheim: Wiley-VCH, 2007.
Dong MG, Wang N. A novel hybrid differential evolution approach to
scheduling of large-scale zero-wait batch processes with setup
times. Comput Chem Eng 2012; 45: 72–83.
dos Santos GS, Luvizotto LGJ, Mariani VC, Coelho LdS. Least
squares support vector machines with tuning based on
chaotic differential evolution approach applied to the
identification of a thermal process. Expert Syst Appl 2012;
39: 4805–4812.
Dragoi EN, Curteanu S, Leon F, Galaction AI, Cascaval D. Modeling of
oxygen mass transfer in the presence of oxygen-vectors using
neural networks developed by differential evolution algorithm.
Eng Appl Artif Intell 2011; 24: 1214–1226.
Dragoi EN, Curteanu S, Lisa C. A neuro-evolutive technique applied
for predicting the liquid crystalline property of some organic
compounds. Eng Optim 2012a; 44: 1261–1277.
Brought to you by | Universita degli Studi di Padova
Authenticated
Download Date | 3/28/16 10:48 AM
Dragoi EN, Curteanu S, Fissore D. Freeze-drying modeling and
monitoring using a new neuro-evolutive technique. Chem Eng
Sci 2012b; 72: 195–204.
Dragoi EN, Curteanu S, Galaction AI, Cascaval D. Optimization
methodology based on neural networks and self-adaptive
differential evolution algorithm applied to an aerobic
fermentation process. Appl Soft Comput 2013a; 13: 222–238.
Dragoi EN, Curteanu S, Fissore D. On the use of artificial neural
networks to monitor a pharmaceutical freeze-drying process.
Drying Technol 2013b; 31: 72–81.
Eiben G, Schut MC. New ways to calibrate evolutionary algorithms. In:
Siarry P, Michalewicz Z, editors. Advances in metaheuristics for
hard optimization. Berlin/Heidelberg: Springer, 2008: 153–177.
Errico M, Pirellas P, Torres-Ortega CE, Rong BG, Segovia-
Hernandez JG. A combined method for the design and
optimization of intensified distillation systems. Chem Eng
Proc: Proc Intensif 2014; 85: 69–76.
Fan Q, Yan X. Differential evolution algorithm with self-adaptive
strategy and control parameters for P-xylene oxidation process
optimization. Soft Comput 2014; 19: 1363–1391.
Feoktistov V. Differential evolution: in search of solutions. Berlin/
Heideberg: Springer, 2006.
Garlapati V, Banerjee R. Optimization of lipase production using
differential evolution. Biotechnol Bioproc Eng 2010; 15: 254–260.
Geraili A, Sharma P, Romagnoli JA. A modeling framework for design
of nonlinear renewable energy systems through integrated
simulation modeling and metaheuristic optimization:
applications to biorefineries. Comput Chem Eng 2014; 61:
102–117.
Gong W, Cai Z, Ling C. ODE: a fast and robust differential evolution
based on orthogonal design. In: Sattar A, Kang Bh, editors. AI
2006: advances in artificial intelligence. Berlin/Heidelberg:
Springer, 2006: 709–718.
Gong W, Cai Z, Jiang L. Enhancing the performance of differential
evolution using orthogonal design method. Appl Math Comput
2008; 206: 56–69.
Greenwood G, Tyrrell A. Introduction to evolvable hardware: a
practical guide for designing self-adaptive systems. New Jersey:
Wiley-IEEE Press, 2006.
Gujarathi AM, Babu BV. Improved multiobjective differential
evolution (MODE) approach for purified terephthalic acid (PTA)
oxidation process. Mater Manuf Proc 2009; 24: 303–319.
Gujarathi AM, Babu BV. Multi-objective optimization of industrial
styrene reactor: adiabatic and pseudo-isothermal operation.
Chem Eng Sci 2010a; 65: 2009–2026.
Gujarathi AM, Babu BV. Hybrid multi-objective differential evolution
(H-MODE) for optimisation of polyethylene terephthalate (PET)
reactor. Int J Bio-Inspired Comput 2010b; 2: 213–221.
Gujarathi AM, Babu BV. Differential evolution strategies for
multi-objective optimization. In: Proceedings of the
International Conference on Soft Computing for Problem
Solving (SocProS 2011). Berlin: Springer, 2012: 63–71.
Gujarathi AM, Babu BV. Multi-objective optimization of low density
polyethylene (LDPE) tubular reactor using strategies of
differential evolution. In: Zelinka I, Sniel V, Abraham A, editors.
Handbook of optimization. Berlin Heidelberg: Springer, 2013:
615–639.
He RJ, Yang ZY. Differential evolution with adaptive mutation and
parameter control using Lévy probability distribution. J Comput
Sci Technol 2012; 27: 1035–1055.
 E.N. Dragoi and S. Curteanu: Differential evolution in chemical engineering      177
Hocalar A, Turker M, Karakuzu C, Yuzgec U. Comparison of different
estimation techniques for biomass concentration in large scale
yeast fermentation. ISA Tran 2011; 50: 303–314.
Holtbruegge J, Kuhlmann H, Lutze P. Process analysis and
economic optimization of intensified process alternatives
for simultaneous industrial scale production of dimethyl
carbonate and propylene glycol. Chem Eng Res Des 2015; 93:
411–431.
Hu C, Yan X. An immune self-adaptive differential evolution
algorithm with application to estimate kinetic parameters for
homogeneous mercury oxidation. Chin J Chem Eng 2009a; 17:
232–240.
Hu C, Yan X. A novel adaptive differential evolution algorithm with
application to estimate kinetic parameters of oxidation in
supercritical water. Eng Optim 2009b; 41: 1051–1062.
Hu C, Yan X. A hybrid differential evolution algorithm integrated
with an ant system and its application. Comput Math Appl
2011; 62: 32–43.
Huang SR, Wu CC, Lin CY, Chen HT. Parameter optimization of
the biohydrogen real time power generating system using
differential evolution algorithm. Int J Hydrogen Energ 2010; 35:
6629–6633.
Iranshahi D, Bahmanpour AM, Paymooni K, Rahimpour MR,
Shariati A. Simultaneous hydrogen and aromatics
enhancement by obtaining optimum temperature profile and
hydrogen removal in naphtha reforming process; a novel
theoretical study. Int J Hydrogen Energ 2011; 36: 8316–8326.
Iranshahi D, Pourazadi E, Paymooni K, Rahimpour MR. Utilizing DE
optimization approach to boost hydrogen and octane number
in a novel radial-flow assisted membrane naphtha reactor.
Chem Eng Sci 2012; 68: 236–249.
Iranshahi D, Rahimpour MR, Paymooni K, Pourazadi E. Utilizing DE
optimization approach to boost hydrogen and octane number,
through a combination of radial-flow spherical and tubular
membrane reactors in catalytic naphtha reformers. Fuel 2013;
111: 1–11.
Kapadi MD, Gudi RD. Optimal control of fed-batch fermentation
involving multiple feeds using differential evolution. Process
Biochem 2004; 39: 1709–1721.
Kazimipour B, Li X, Qin AK. Initialization methods for large scale
global optimization. In: 2013 IEEE Congress on Evolutionary
Computation (CEC). IEEE 2013: 2750–2757.
Khademi MH, Setoodeh P, Rahimpour MR, Jahanmiri A. Optimization
of methanol synthesis and cyclohexane dehydrogenation in
a thermally coupled reactor using differential evolution (DE)
method. Int J Hydrogen Energ 2009; 34: 6930–6944.
Khademi MH, Rahimpour MR, Jahanmiri A. Differential evolution (DE)
strategy for optimization of hydrogen production, cyclohexane
dehydrogenation and methanol synthesis in a hydrogen-
permselective membrane thermally coupled reactor. Int J
Hydrogen Energ 2010; 35: 1936–1950.
Khademi MH, Farsi M, Rahimpour MR, Jahanmiri A. DME synthesis
and cyclohexane dehydrogenation reaction in an optimized
thermally coupled reactor. Chem Eng Proc: Proc Intensif 2011;
50: 113–123.
Koch K, Sudhoff D, Kreis S, Gorak A, Kreis P. Optimisation-based
design method for membrane-assisted separation processes.
Chem Eng Proc: Proc Intensif 2013; 67: 2–15.
Kumar S, Datta D, Babu BV. Estimation of equilibrium parameters
using differential evolution in reactive extraction of propionic
Brought to you by | Universita degli Studi di Padova
Authenticated
Download Date | 3/28/16 10:48 AM
acid by tri-n-butyl phosphate. Chem Eng Proc: Proc Intensif
2011; 50: 614–622.
Lahiri SK, Ghanta KC. The support vector regression with the
parameter tuning assisted by a differential evolution
technique: study of the critical velocity of a slurry flow in a
pipeline. Chem Ind Chem Eng Quaterly 2008; 14: 191–203.
Lahiri SK, Ghanta KC. Support vector regression with parameter
tuning assisted by differential evolution technique: study on
pressure drop of slurry flow in pipeline. Korean J Chem Eng
2009a; 26: 1175–1185.
Lahiri SK, Ghanta KC. Artificial neural network model with the
parameter tuning assisted by a differential evolution
technique: the study of the hold up of the slurry flow in a
pipeline. Chem Ind Chem Eng Quarterly 2009b; 15: 103–117.
Lahiri SK, Khalfe NM. Novel soft sensor modeling and process
optimization technique for commercial petrochemical plant.
Asia-Pacific J Chem Eng 2010; 5: 721–731.
Lahiri SK, Khalfe N, Garawi MA. Process modeling and optimization
strategies integrating neural networks and differential
evolution. Hydrocarbon Proc 2008; 87: 35–50.
Lei, P, Wang Y, Dai G. UDE: differential evolution with uniform
design. In: 2010 Third International Symposium on Parallel
Architectures, Algorithms and Programming (PAAP), 2010:
239–246.
Levy A, College B, Maine B. The basics of practical optimization.
Philadelphia: Society for Industrial and Applied Mathematics
(SIAM), 2009.
Li S, Li Y. A New Self-adaption differential evolution algorithm
based component model. In: Cai Z, Hu C, Kang Z, Liu Y, editors.
Advances in computation and intelligence. Berlin/Heidelberg:
Springer, 2010: 54–63.
Lin C, Qing A, Feng Q. A comparative study of crossover in
differential evolution. J Heuristics 2010; 17: 675–703.
Liu PK, Wang FS. Hybrid differential evolution with geometric mean
mutation in parameter estimation of bioreaction systems with
large parameter search space. Comput Chem Eng 2009; 33:
1851–1860.
Liu PK, Wang FS. Hybrid differential evolution including geometric
mean mutation for optimization of biochemical systems.
J Taiwan Instit Chem Eng 2010; 41: 65–72.
Liu B, Reckhow DA, Li Y. A two-site chlorine decay model for the
combined effects of pH, water distribution temperature and
in-home heating profiles using differential evolution. Water
Res 2014; 53: 47–57.
Luo L, Zhang L, Gu X. A hybrid differential evolution algorithm
with invasive weed optimization and its application to
modeling of carbon content. In: 2014 IEEE International
Conference on System Science and Engineering (ICSSE). IEEE,
2014: 76–81.
Mallipeddi R, Suganthan PN, Pan QK, Tasgetiren MF. Differential
evolution algorithm with ensemble of parameters and mutation
strategies. Appl Soft Comput 2011; 11: 1679–1696.
Mariani VC, Barbosa de Lima AG, dos Santos Coelho L. Apparent
thermal diffusivity estimation of the banana during drying
using inverse method. J Food Eng 2008; 85: 569–579.
Mareci D, Dragoi EN, Bolat G, Chelariu R, Gordin D, Curteanu S.
Modelling the influence of pH, fluoride, and caffeine on the
corrosion resistance of TiMo alloys by artificial neural networks
developed with differential evolution algorithm. Mater Corros
2015; 66: 982–994.
178      E.N. Dragoi and S. Curteanu: Differential evolution in chemical engineering
Mavrovouniotis M, Yang S. Ant colony optimization with immigrants
schemes for the dynamic travelling salesman problem with
traffic factors. Appl Soft Comput 2013; 13: 4023–4037.
Mekapati S, Gade PR. Differential evolution with tabu list for global
optimization: evaluation of two versions on benchmark and
phase stability problems. In: Anonymous stochastic global
optimization. Singapore: World Scientific, 2010: 465–504.
Mendes R, Rocha I, Pinto J, Ferreira E, Rocha M. Differential
evolution for the offline and online optimization of fed-batch
fermentation processes. In: Chakraborty U, editor. Advances
in differential evolution. Berlin/Heidelberg: Springer, 2008:
299–317.
Miranda-Galindo EY, Segovia-Hernandez JG, Hernandez-Castro S,
Bonilla-Petriciolet A, Rangaiah GP. Multiobjective optimization
in distillation with reactor-side for hydrodesulfurization
process of diesel. Comput Aided Chem Eng 2012; 30: 682–686.
Mohamed AW, Sabry HZ, Abd-Elaziz T. Real parameter optimization
by an effective differential evolution algorithm. Egypt Inf J
2013; 14: 37–53.
Montana D, VanWyk E, Brinn M, Montana J, Milligan S. Evolution of
internal dynamics for neural network nodes. Evol Intell 2009;
1: 233–251.
Moonchai S, Madlhoo W, Jariyachavalit K, Shimizu H, Shioya S,
Chauvatcharin S. Application of a mathematical model and
differential evolution algorithm approach to optimization of
bacteriocin production by Lactococcus lactis C7. Bioprocess
Biosyst Eng 2005; 28: 15–26.
Nian X, Wang Z, Qian F. A Hybrid algorithm based on differential
evolution and group search optimization and its application on
ethylene cracking furnace. Chin J Chem Eng 2013; 21: 537–543.
Nobakhti A, Wang H. A simple self-adaptive differential evolution
algorithm with application on the ALSTOM gasifier. Appl Soft
Comput 2008; 8: 350–370.
Nourbakhsh H, Yazdizadeh M, Esmaeilzadeh F. Prediction of
asphaltene precipitation by the extended Flory-Huggins model
using the modified Esmaeilzadeh-Roshanfekr equation of
state. J Petr Sci Eng 2011; 80: 61–68.
Ortegaa CET, Hernandeza JGS, Gomez FI, Castroa SH, Rongb BG,
Erricob M. Design and optimization of a process based on
extractive distillation for the sweetening natural gas. Chem Eng
Proc: Proc Intensif 2013; 74: 55–68.
Pant M, Ali M, Abraham A. Mixed mutation strategy embedded
differential evolution. In: IEEE Congress on Evolutionary
Computation (CEC′09). IEEE 2009a: 1240–1246.
Pant M, Thangaraj R, Abraham A, Grosan C. Differential evolution
with Laplace mutation operator. In: Proceedings of the Eleventh
Conference on Congress on Evolutionary Computation. IEEE
Press, 2009b: 2841–2849.
Parvasi P, Khosravanipour Mostafazadeh A, Rahimpour MR. Dynamic
modeling and optimization of a novel methanol synthesis loop
with hydrogen-permselective membrane reactor. Int J Hydrogen
Energ 2009; 34: 3717–3733.
Peng L, Wang Y. Differential evolution using unifor-quasi-
opposition for initializing the population. Inf Technol J 2010; 9:
1629–1634.
Pirdashti M, Movagharnejad K, Curteanu S, Dragoi EN, Rahimpour F.
Prediction of partition coefficients of guanidine hydrochloride
in PEG-phosphate systems using neural networks developed
with differential evolution algorithm. J Ind Eng Chem. 2015; 27:
268–275.
Brought to you by | Universita degli Studi di Padova
Authenticated
Download Date | 3/28/16 10:48 AM
Pourazadi E, Vakili R, Iranshahi D, Jahanmiri A, Rahimpour MR.
Optimal design of a thermally coupled fluidised bed heat
exchanger reactor for hydrogen production and octane
improvement in the catalytic naphtha reformers. Can J Chem
Eng 2013; 91: 54–65.
Precup RE, Tomescu ML, Radac MB, Petriu EM, Preitl S, Dragoiş CA.
Iterative performance improvement of fuzzy control systems for
three tank systems. Expert Sys Appl 2012; 39: 8288–8299.
Price K, Storn R, Lampinen J. Differential evolution. A practical
approach to global optimization. Berlin: Springer, 2005.
Qian F, Xu B, Qi R, Tianfield H. Self-adaptive differential evolution
algorithm with e constrained-domination principle for
constrained multi-objective optimization. Soft Comput 2012;
16: 1353–1372.
Qin AK, Huang VL, Suganthan PN. Differential evolution algorithm
with strategy adaptation for global numerical optimization.
IEEE Trans Evol Comput 2009; 13: 398–417.
Qin H, Zhou J, Lu Y, Wang Y, Zhang Y. Multi-objective differential
evolution with adaptive Cauchy mutation for short-term
multi-objective optimal hydro-thermal scheduling. Energ
Convers Manag 2010; 51: 788–794.
Rahimpour MR, Parvasi P, Setoodeh P. Dynamic optimization of a
novel radial-flow, spherical-bed methanol synthesis reactor
in the presence of catalyst deactivation using differential
evolution (DE) algorithm. Int J Hydrogen Energ 2009; 34:
6221–6230.
Rahimpour MR, Iranshahi D, Bahmanpour AM. Dynamic optimization
of a multi-stage spherical, radial flow reactor for the naphtha
reforming process in the presence of catalyst deactivation
using differential evolution (DE) method. Int J Hydrogen Energ
2010; 35: 7498–7511.
Rahimpour MR, Mirvakili A, Paymooni K. Differential evolution
(DE) strategy for optimization of hydrogen production and
utilization in a thermally coupled membrane reactor for
decalin dehydrogenation and Fischer-Tropsch synthesis in GTL
technology. Int J Hydrogen Energ 2011; 36: 4917–4933.
Rahimpour MR, Iranshahi D, Pourazadi E, Bahmanpour AM.
Boosting the gasoline octane number in thermally coupled
naphtha reforming heat exchanger reactor using de
optimization technique. Fuel 2012; 97: 109–118.
Rahimpour MR, Mirvakili A. A novel configuration of decalin and
hydrogen loop in optimized thermally coupled reactors in GTL
technology via differential evolution method. J Ind Eng Chem
2013; 19: 508–522.
Rahman I, Bokade VV, Patnaik A, Das AK. Optimisation of
ethylbenzene conversion and selectivity of p-diethylbenzene
using differential evolution technique via OLS modelling.
Indian Chem Eng 2010; 52: 87–98.
Rahnamayan S, Tizhoosh H. differential evolution via exploiting
opposite populations. In: Tizhoosh H, Ventresca M, editors.
Oppositional concepts in computational intelligence. Berlin/
Heidelberg: Springer, 2008: 143–160.
Rocha M, Pinto J, Rocha I, Ferreira E. Evaluating evolutionary
algorithms and differential evolution for the online
optimization of fermentation processes. In: Marchiori E,
Moore J, Rajapakse J, editors. Evolutionary computation,
machine learning and data mining in bioinformatics. Berlin/
Heidelberg: Springer, 2007: 236–246.
Safe M, Khazraee SM, Setoodeh P, Jahanmiri AH. Model reduction
and optimization of a reactive dividing wall batch distillation
 E.N. Dragoi and S. Curteanu: Differential evolution in chemical engineering      179
column inspired by response surface methodology and
differential evolution. Math Comput Modell Dynam Syst 2013;
19: 29–50.
Salman A, Engelbrecht AP, Omran MGH. Empirical analysis of
self-adaptive differential evolution. Eu J Oper Res 2007; 183:
785–804.
Samimi F, Kabiri S, Mirvakili A, Rahimpour MR. Simultaneous
dimethyl ether synthesis and decalin dehydrogenation in an
optimized thermally coupled dual membrane reactor. J Nat Gas
Sci Eng 2013a; 14: 77–90.
Samimi F, Bayat M, Karimipourfard D, Rahimpour MR, Keshavarz P.
A novel axial-flow spherical packed-bed membrane reactor for
dimethyl ether synthesis: simulation and optimization. J Nat
Gas Sci Eng 2013b; 13: 42–51.
Samimi F, Kabiri S, Mirvakili A, Rahimpour MR. The concept of
integrated thermally double coupled reactor for simultaneous
production of methanol, hydrogen and gasoline via differential
evolution method. J Nat Gas Sci Eng 2013c; 14: 144–157.
Santos KG, Lobato FS, Lira TS, Murata VV, Barrozo MAS. Sensitivity
analysis applied to independent parallel reaction model for
pyrolysis of bagasse. Chem Eng Res Des 2012; 90: 1989–1996.
Sharma S, Rangaiah GP. A hybrid multi-objective optimization
algorithm. In: 5th International Symposium on Design,
Operation, and Control of Chemical Processes, 2010:
1494–1503.
Sharma S, Rangaiah GP. Multi-objective optimization of a bio-diesel
production process. Fuel 2013a; 103: 269–277.
Sharma S, Rangaiah GP. An improved multi-objective differential
evolution with a termination criterion for optimizing chemical
processes. Comput Chem Eng 2013b; 56: 155–173.
Shi Y, Zhong X. Hierarchical differential evolution for parameter
estimation in chemical kinetics. In: Ho TB, Zhou ZH, editors.
PRICAI 2008: Trends in artificial intelligence. Berlin/
Heidelberg: Springer, 2008: 870–879.
Snyman J. Practical mathematical optimization. An introduction
to basic optimization theory and classical and new
gradient-based algorithms. New York: Springer, 2005.
Storn, R. On the usage of differential evolution for function
optimization. In: 1996 Biennial Conference of the North
American Fuzzy Information Processing Society. NAFIPS 1996:
519–523.
Storn R, Price K. Differential evolution – a simple and efficient
adaptive scheme for global optimization over continuous
spaces. TR-95-012. 1995. Berkley.
Sudheer NVVS, Pavan KK. Effect of carburizing flame and oxidizing
flame on surface roughness in turning of aluminium metal
matrix composite and differential evolution optimization of
process parameters. Procedia Mater Sci 2014; 6: 840–850.
Tan KC, Goh CK, Mamun AA, Ei EZ. An evolutionary artificial immune
system for multi-objective optimization. Eu J Oper Res 2008;
187: 371–392.
Thangaraj R, Pant M, Abraham A. New mutation schemes for
differential evolution algorithm and their application to the
optimization of directional over-current relay settings. Appl
Math Comput 2010; 216: 532–544.
Thangaraj R, Pant M, Chelliah TR, Abraham A. Opposition based
chaotic differential evolution algorithm for solving global
optimization problems. 2012 Fourth World Congress on
Nature and Biologically Inspired Computing (NaBIC). IEEE:
2012: 1–7.
Brought to you by | Universita degli Studi di Padova
Authenticated
Download Date | 3/28/16 10:48 AM
Tvrdik J. Adaptive differential evolution and exponential crossover.
In: International Multiconference on Computer Science and
Information Technology IMCSIT 2008. Wisia, 2008: 927–931.
Vakili R, Eslamloueyan R. Optimal design of an industrial scale
dual-type reactor for direct dimethyl ether (DME) production
from syngas. Chem Eng Proc: Proc Intensif 2012; 62: 78–88.
Vakili R, Setoodeh P, Pourazadi E, Iranshahi D, Rahimpour MR.
Utilizing differential evolution (DE) technique to optimize
operating conditions of an integrated thermally coupled direct
DME synthesis reactor. Chem Eng J 2011; 168: 321–332.
Vakili R, Rahimpour MR, Eslamloueyan R. Incorporating differential
evolution (DE) optimization strategy to boost hydrogen
and DME production rate through a membrane assisted
single-step DME heat exchanger reactor. J Nat Gas Sci Eng
2012; 9: 28–38.
Vázquez-Ojeda M, Segovia-Hernández JG, Hernández S, Hernández-
Aguirre A, Kiss AA. Optimization of an ethanol dehydration
process using differential evolution algorithm. Comput Aided
Chem Eng 2013; 32: 217–222.
Wang FS, Cheng WM. Simultaneous optimization of feeding rate and
operation parameters for fed-batch fermentation processes.
Biotechnol Progr 1999; 15: 949–952.
Wang C, Gao JH. A differential evolution algorithm with cooperative
coevolutionary selection operation for high-dimensional
optimization. Optim Let 2014; 8: 477–492.
Wang FS, Su TL, Jang HJ. Hybrid differential evolution for problems
of kinetic parameter estimation and dynamic optimization of
an ethanol fermentation process. Ind Eng Chem Res 2001; 40:
2876–2885.
Wang Y, Cai Z, Zhang Q. Enhancing the search ability of differential
evolution through orthogonal crossover. Inf Sci 2012; 185:
153–177.
Wu Y, Lu J, Sun Y. An improved differential evolution for optimization
of chemical process. Chin J Chem Eng 2008; 16: 228–234.
Xu W, Zhang L, Gu X. Modeling of ammonia conversion rate in
ammonia synthesis based on a hybrid algorithm and least
squares support vector regression. Asia-Pacific J Chem Eng
2012; 7: 150–158.
Xu B, Qi R, Zhong W, Du W, Qian F. Optimization of p-xylene
oxidation reaction process based on self-adaptive multi-
objective differential evolution. Chem Intellig Lab Syst 2013;
127: 55–62.
Yacoub RR, Bambang RT, Harsoyo A, Sarwono J. DSP Implementation
of combined FIR-functional link neural network for active noise
control. Int J Artificial Intell 2014; 12: 36–47.
Yazdizadeh M, Rahmani F, Forghani A. Thermodynamic modeling
of CO2 solubility in ionic liquid ([Cn-mim] [Tf2N]; n = 2, 4, 6, 8)
with using Wong-Sandler mixing rule, Peng-Rabinson equation
of state (EOS) and differential evolution (DE) method. Korean J
Chem Eng 2011; 28: 246–251.
Yuzgec U. Performance comparison of differential evolution
techniques on optimization of feeding profile for an industrial
scale baker’s yeast fermentation process. ISA Tran 2010; 49:
167–176.
Zaharie, D. Critical values for the control parameters of differential
evolution algorithms. In: Osmera P, Matousek R, editors.
Proceedings of MENDEL 2002. 8th International Conference on
Soft Computing, 2002: 62–67.
Zamuda A, Brest J. Population reduction differential evolution with
multiple mutation strategies in real world industry challenges.
180      E.N. Dragoi and S. Curteanu: Differential evolution in chemical engineering

In: Swarm and evolutionary computation. Berlin Heidelberg: Silvia Curteanu

Springer, 2012: 154–161.
Zhang J, Sanderson AC. Adaptive differential evolution: a robust
approach to multimodal problem optimization. Berlin:
Springer, 2009.
Zhang H, Bonilla-Petriciolet A, Rangaiah GP. A review on global
optimization methods for phase equilibrium modeling and
calculations. Open Thermodynam J 2011a; 5: 71–92.
Zhang H, Fernandez-Vargas JA, Rangaiah GP, Bonilla-Petriciolet A,
Segovia-Hernandez JG. Evaluation of integrated differential
evolution and unified bare-bones particle swarm optimization
for phase equilibrium and stability problems. Fluid Phase
Equilib 2011b; 310: 129–141.
Zhang H, Rangaiah GP, Bonilla-Petriciolet A. Integrated differential
evolution for global optimization and its performance for
modeling vapor-liquid equilibrium data. Ind Eng Chem Res
2011c; 50: 10047–10061.
Zhang H, Rangaiah GP. One-step approach for heat exchanger
network retrofitting using integrated differential evolution.
Comput Chem Eng 2013; 50: 92–104.
Zhao C, Xu Q, Lin S, Li X. Hybrid differential evolution for estimation
of kinetic parameters for biochemical systems. Chin J Chem
Eng 2013; 21: 155–162.
Zielinski K, Laur R. Stopping criteria for differential evolution in
constrained single-objective optimization. In: Chakraborty U,
editor. Advances in differential evolution. Berlin/Heidelberg:
Springer, 2008: 111–138.
Faculty of Chemical Engineering and
Environmental Protection, Department of
Chemical Engineering/Applied Informatics,
“Gheorghe Asachi” Technical University Iasi,
Bd. Prof.dr.doc. Dimitrie Mangeron, No. 71A,
700050, Iasi, Romania,
silvia_curteanu@yahoo.com
Silvia Curteanu has been a Professor and PhD supervisor in
Chemical Engineering since 2005 at “Gheorghe Asachi” Technical
University of Iasi, Romania, Faculty of Chemical Engineering and
Environmental Protection. She is coordinator of the Applied Infor-
matics Laboratory and Research Center “Chemical and Biochemical
Process Engineering and Advanced Materials”. Her professional
experience and research interests are artificial intelligence tools
applied in chemical engineering, neural networks methodologies
used for modeling purposes, and evolutionary algorithms (genetic,
differential evolution, artificial immune algorithms) applied for
process optimization. Prof. Curteanu has more than 150 publica-
tions (scientific papers and books). Web page address: http://www.
ch.tuiasi.ro/cv/ic/curteanusilvia/index.html.

Bionotes

Elena Niculina Dragoi

Faculty of Chemical Engineering and
Environmental Protection, Department of
Chemical Engineering/Applied Informatics,
“Gheorghe Asachi” Technical University Iasi,
Bd. Prof.dr.doc. Dimitrie Mangeron, No. 71A,
700050, Iasi, Romania

Elena Niculina Dragoi received a BA degree in Computer Science and a

MA degree in Distributed Systems from the “Gheorghe Asachi” Univer-
sity of Iasi, Faculty of Automatic Control and Computer Engineering, in
2008 and 2009, respectively, and a PhD degree in Chemical Engineer-
ing from the “Gheorghe Asachi” University of Iasi, Faculty of Chemical
Engineering and Environmental Protection, in 2012. Her research
interests include evolutionary computation, bio-inspired algorithms,
artificial neural networks, and their application for solving modeling
and optimization problems related to chemical engineering.

Brought to you by | Universita degli Studi di Padova

Authenticated
Download Date | 3/28/16 10:48 AM
Copyright of Reviews in Chemical Engineering is the property of De Gruyter and its content
may not be copied or emailed to multiple sites or posted to a listserv without the copyright
holder's express written permission. However, users may print, download, or email articles for
individual use.