MPZ 5230 Model Answer 04

Academic Year 2009

1 - population mean filling time of machine 1.
2 - Population mean filling time of machine 2.
3 - Population mean filling time of machine 3.

(1)

x1 =

Hypothesis:

H o : 1 = 2 = 3 (Null hypothesis)
H1 : 1 , 2 and 3 are not all equal(Alternative hypothesis)

Machine 1

Machine 2

25.4 + 26.31 + 24.10 + 23.74 + 25.10

5

23.4 + 21.8 + 23.5 + 22.75 + 21.6

5

x2 =

= 24.93

(x

x1

(x

0.47
1.38
-0.83
-1.19
0.17

S =
2
1

x1

(x

0.2209
1.9044
0.6889
1.4161
0.0289
1

x1

(x

x1

n 1

(n

x2

4.2592
=
4

1)

( nT K )

20 + 22.2 + 19.75 + 20.6 + 20.4

5

S =

x2

0.6241
0.6561
0.7921
0.0196
1.0201

(x
2
2

= 20.59

(x

0.79
-0.81
0.89
0.14
-1.01

= 1.0648

2 =

= 4.2592

x3 =

= 22.61

(x

Machine 3

x2

(x

x2

(x

x3

(x

n 1
= 0.778

3.112
4

(x

-0.59
1.61
-0.84
0.01
-0.19

= 3.112

S32 =

x3

0.3481
2.5921
0.7056
0.0001
0.0361
3

x3

(x

x3

n 1

= 3.682

) = 3.682

= 0.9205

S2j (within column variance)

5 1
[1.0648 + 0.778 + 0.9205] = 0.9211
15 3

x = Grand Mean of machine 1, 2 & 3 =

1
( 24.93 5 + 22.61 5 + 20.59 5) = 22.71
15

xx

( )

5
5
5

24.93
22.61
20.59

22.71
22.71
22.71

2.22
-0.1
-2.12

4.9284
0.01
4.4944

nj xj x

2 =

nj xj x
k 1

Fcal =

xx

( )

n xx

24.642
0.05
22.472

= 47.164

47.164
= 23.582
2

(between column variance)

23.582
= 25.6
0.9211

Ftab = F2,12 = 3.89

0.05 of area
calculated value

3.89

Ftab < Fcal

25.6

Hence we can reject Ho

Therefore there is evidence that the mean filling times of machine 1, 2 and 3 are
different under a 0.05 level of significance.
(2)

(a)

SM - Population mean of no. of years of education for Sinhalese male.

SF - Population mean of no. of years of education for Sinhalese female.
TM - Population mean of no. of years of education for Tamil male.
TF - Population mean of no. of years of education for Tamil female.

Hypothesis:

H o : SM = SF = TM = TF

H1 : SM , SF , TM and TF are not all equal

(Null hypothesis)
(Alternative hypothesis)

(b) No.
( c) Analysis of variance technique is important to test for differences among several
sample means. Compared with using multiple t-tests, one-way and two-way ANOVA
require fewer measurements to discover significant effects. This is one reason why

ANOVA is used frequently when analyzing data from statistically designed

experiments.
(d)
Sinhalese Male

x SM =

Sinhalese Female

= 15.67

(x

SM

x SM

) (x
2

SF x SF

(x

SF

x SF

16-15.69=0.33

0.1089

16-12.5=3.5

12.25

18-15.67=2.33

5.4289

12-12.5=-0.5

0.25

14-15.67=-1.67

2.7889

11-12.5=-1.5

2.25

14-15.67=-1.67

2.7889

14-12.5=1.5

2.25

16-15.67=0.33

0.1089

11-12.5=-1.5

2.25

16-15.67=0.33

0.1089

11-12.5=-1.5

2.25

(x
2
SM

x SM = 11.3334

SM

(x

SM

x SM

11.3334
5

2 =
=

x SF

(x

S =
2
SF

SF

x SF

21.5
5

nT k

= 12.25

(x

TM

x TM

(x

6.76

14-12.25=1.75

12-13.4=-1.4

1.96

10-12.25=-2.25

14-13.4=0.6

0.36

12-12.25=-0.25

14-13.4=0.6

0.36

13-12.25=0.75

11-13.4=-2.4

5.76

(x

TM

x TM

2
TM

(x

TM

x TM

TF

15.2
4

= 3.8

S2j (within column variance)

6 1
6 1
5 1
4 1
2.2667 +
4.3 +
3.8 +
2.9167
21 4
21 4
21 4
21 4
1
56.7836
= 3.3402
{11.3335 + 21.5 + 15.2 + 8.7501} =
17
17

1
285.02
= 13.57
(15.67 6 + 12.5 6 + 13.4 5 + 12.25 4 ) =
21
21

2
TF

(x

x TF

5.0625
0.0625
0.5625

TF

= 8.75

x TF

n 1

TF

3.0625

x TF

n 1
=

(x

= 15.2

x = Grand mean of four groups

x TF x TF

16-13.4=2.6

= 4.3

n j 1

x TM x TM

= 21.5

n 1

= 2.2667

2 =

SF

n 1

(x

14 + 10 + 12 + 13
4

x TF =

= 13.4

= 12.5
x SM x SM

Tamil Female

16 + 12 + 11 + 14 + 11 + 11
16 + 12 + 14 + 14 + 11
x TM =
6
5

x SF =

16 + 18 + 14 + 16 + 16
6

Tamil Male

8.75
3

= 2.9167

xx

( )

15.67

13.57

2.1

4.41

26.46

12.5

13.57

-1.07

1.1449

6.8694

13.4

13.57

-0.17

0.0289

0.1445

12.25

13.57

-1.32

1.7424

6.9696

nj xj x

( )

n xx

= 40.4435

nj xj x

2 =

Fcal =

xx

k 1

40.4435
= 13.48
3

(Between column variance)

13.48
= 4.0357
3.3402

Ftab = F3,17 = 3.20

Accepted region

0.05 of area
calculated value

Ftab < Fcal

3.2

4.0357

Therefore we can reject Ho

We can conclude that there is a evidence of difference between the four groups at 0.05
level of significance.
(3)(a) Simulation is
-

Designing a model for a real or proposed system

Executing the model on a computer

Analyzing the execution output

Difference from analytical solutions

Approximate answer to an exact problem is better than the exact

answer to an approximate problem.

(b) Advantages of using simulation:

Helps in the understanding of complex problems or systems.

Permits study of alternatives under controlled conditions.

Avoids risk and disruptive experiments with actual system

Compresses time to reveal long range effects

Is less costly than experiments with real-world system

(c) Monte Carlo Simulation refers to the use of random sampling to estimate the output of an
experiment. It is regarded as the forerunner to present-day simulation.
(d) Data are given in frequencies. Then formulate a probability distribution and Cumulative
Probability distribution.
Demand

Frequency

Probability

Cumulative

(tons/day)

(No. days)

P(x)

probability

10

6
= 0.10
60

0.10

11

18

18
= 0.30
60

0.40

12

15

15
= 0.25
60

0.65

13

12

12
= 0.20
60

0.85

6
= 0.10
60

0.95

3
= 0.05
60

1.00

14
15

Total = 60
Next assign random number intervals, so that the number of values available to each class
corresponds with the probability.
Cumulative
Demand

P(x)

probability
F(x)

Corresponding
random numbers

10

0.10

0.10

00 09

11

0.30

0.40

10 39

12

0.25

0.65

40 64

13

0.20

0.85

65 84

14

0.10

0.95

85 94

15

0.05

1.00

95 - 99

Choose two-digit random numbers (00 to 99) from the table of random numbers in
your text book.

(4)

Demand

Random

Simulated

Period

Number

Demand

39

11

65

13

76

13

45

12

45

12

19

11

90

14

Let x1 the number of tractor backhoes used

x 2 - the number of hydraulic backhoes used

Objective Function:

Minimize

Constraints;

200x1 + 1000x 2 10, 000

200x1 1600
1000x 2 3000

x1 + x 2 12

z = 400x1 + 1110x 2

x2

( 0,12 )

x 1 + x 2 = 12

( 0,10 )

Feasible
region

C
( 0,3 )
A

200 x 1+ 100 x 2 = 10000

( 8,0 ) ( 12,0 )

( 50,0 )

200 x 1= 1600

Corner feasible points:

A = (8, 3)

objective function = 400 8 + 1110 3 = 6530

B = (9, 3)

objective function = 6930

C = (8, 4)

objective function = 7640

Optimal solution given by the minimum objective function

Optimal solution:

x1 = 8

x2 = 3

Minimum daily operating cost = 6530

(5)

Minimize

z = 4x1 + x 2

Subject to

3x1 + x 2 = 3

1000 x 2 = 3000

4x1 + 3x 2 6
x1 + 2x 2 4
x1 , x 2 0

r = a1 + a 2 --------- (1)

Phase I

Minimize

Subject to

3x1 + x 2 + a1 = 3 ----------- (2)

4x1 + 3x 2 S1 + a 2 = 6 ---------- (3)

x1 + 2x 2 + S2 = 4 ---------- (4)

Table 1:

Pivot element
x1

Basic

a1

x2

S1

S2

a1

a2

RHS

ratio

3
=1
3

a2

-1

6 3
=
4 2

S2

4
=4
1

-4

-1

r1

-1

To eliminate a1 and a 2 from equation (1), we add the constraints (2) and (3), then
subtract the result from (1).
Then we obtain r1 ( 3 3x1 x 2 ) ( 6 4x1 3x 2 + S1 ) = 0

r1 + 7x1 + 4x 2 S1 = 9
Table 2:
Basic

x1

x2

S1

S2

a1

a2

RHS

ratio

x1

1
3

1
3

1
=3
1
3

5
3

-1

4
3

2 6
=
5
5
3

1
3

3 9
=
5
5
3

a2

S2

5
3

1
3

r1

5
3

-1

4
3

7
3

Table 3:
Basic

x1

x2

S1

S2

a1

a2

x1

1
5

3
5

x2

S2

1
5

r1

3
5

1
5

ratio

3
5

3
5

6
5

-1

3
5

1
5

18
5

-1

-1

4
5

RHS

At this point all the artificial variables are non-basic variables and have value zero.
Phase II
(1)

Delete the columns a1 and a 2

(2)

Delete the auxiliary objective function row.

After deleting the artificial columns, the original problem is written as,
Minimize z = 4x1 + x 2
Subject to

1
3
x1 + S1 =
5
5

3
6
x 2 S3 =
5
5
S1 + S2 = 1

x1 , x 2 ,S1 ,S2 0
Basic variable x1 , x 2 have non-zero coefficients in the z row, they must be substituted
out.
Now z row = old z row + (4 x1 row) + (1 x2 row)
= 0 0

1
18
0
5
5

This is same as that available in the final tableau of phase I.

We have the initial tableau for phase II.
Basic

x1

x2

S1

S2

RHS

ratio

x1

1
5

3
5

x2

6
5

-3

S2

1
5

18
5

Basic

x1

x2

S1

S2

RHS

x1

x2

3
5

9
5

S1

3
5

1
5

1
5

ratio

2
5

17
5

Since the objective row in above table has no positive entries, we have found an
optimal solution:
2
9
x1 = , x 2 = , S1 = 1
5
5
z=

(6)

(a)

17
5
Primal problem
x1 no. of x1 products
x 2 no. of x2 products

Objective function

Maximize z = 12x1 + 4x 2

Subject to
3x1 + 2x 2 6

(Machine M1)

1
x1 + x 2 4
2

(Machine M2)

x1 , x 2 0

(b)

Dual problem:
Objective function
Subject to

Minimize z = 6y1 + 4y 2

1
3y1 + y 2 12
2

2y1 + y 2 4
y1 , y 2 0

(c)

Standard form
Subject to

Minimize

z = 6y1 + 4y 2

1
3y1 + y 2 S1 + a1 = 12
2
2y1 + y 2 S2 + a 2 = 4 2

y1 , y 2 , a1 , a 2 ,S1 ,S2 0
We can use use two-phase method.
Phase I:
Minimize

r = a1 + a 2

Subject to

3y1 + y 2 S1 + a1 = 12
2y1 + y 2 S2 + a 2 = 4
y1 , y 2 , a1 , a 2 ,S1 ,S2 0

1
r1 12 3y1 y 2 + S1 ( 4 2y1 y 2 + S2 ) = 0
2
New row
3
r1 + 5y1 + y 2 S1 S2 = 16
2

Table 1:
Basic

y1

y2

S1

S2

a1

a2

RHS

ratio

a1

0.5

-1

12

a2

-1

-6

-4

r1

1.5

-1

-1

16

Basic

y1

y2

S1

S2

a1

a2

RHS

ratio

a1

-1

-1

1.5

-1.5

y1

0.5

-0.5

0.5

-1

-3

12

-1

-1

1.5

-2.5

Basic

y1

y2

S1

S2

a1

a2

RHS

S2

2
3

2
3

-1

y1

1
6

1
3

1
3

-3

-2

24

-1

-1

Table 2:

Table 3:

2
3

At this point all the artificial variables are non-basic variables & have value zero.

ratio

Phase II
(1)

Delete the columns a1 & a 2

(2)

Delete the auxiliary objective function row.

New z row = old z row + (6 y1 row)

0 -3 -2 0 24
This is the initial tableau for phase II.
Basic

y1

y2

S1

S2

RHS

S2

2
3

y1

1
6

1
3

-3

-2

24

2
3

ratio

Since the objective row in the above table has no positive entries, the optimal solution is
y1 = 4, S2 = 4, z=24

The optimal values of the primal variables can be obtained from the dual optimal solution.
Optimal values of
Primal variables

coefficients of basic
optimal solution of the basic
=
variables in the initial
var iables of the dual problem
tabelau of dual problem

Using the above table,

( x1

x 2 ) = ( S2 y1 )

1 2

= ( 2 0)

Solution for the primal problem is :

x1 = 2, x 2 = 0, z = 24