You are on page 1of 2

Algorithm Design Techniques Due: Tuesday 7th October

Assignment 2: Linear Programming

(1) Standard Form.


(a) Write the following linear programming problem in standard form.

min 5x1 − 7x2 + 2x3 − x4


s.t. 8x1 + x2 − x3 − 3x4 ≤ 6
2x2 + x4 ≥ 10
−x1 + 3x3 + x4 ≤ −3
x2 − 5x3 − 2x4 = 0
x1 , x3 , x4 ≥ 0

(b) Give the dual problem for the linear programming problem from part a).
[Use the usual dual format, that is, min{bT y : AT y ≥ c, y ≥ 0}.]
(c) Convert this dual problem into standard form.

(2) Feasibility.
(a) Give necessary and sufficient conditions on the numbers a1 and a2 so that
the following LP:
(i) has an optimal solution, (ii) is infeasible (iii) is unbounded.

max x1 + x2
s.t. a1 x 1 + a2 x 2 ≤ 1
x1 , x2 ≥ 0

(b) Repeat the question for the following LP.

max x1 + x2
s.t. a1 x1 + a2 x2 ≤ −1
x1 , x2 ≥ 0

1
2

(3) Geometry. Plot the feasible region for the following LP and identify the opti-
mal solution.
max 5x1 + 3x2
s.t. 5x1 − 2x2 ≥ 0
x1 + x2 ≤ 7
x1 ≤ 5
x1 , x2 ≥ 0

(4) The Simplex Method.


(a) Take the dual of the primal
max 5x1 + 6x2 + 9x3 + 8x4
s.t. x1 + 2x2 + 3x3 + x4 ≤ 5
x1 + x2 + 2x3 + 3x4 ≤ 3
x1 , x2 , x3 , x4 ≥ 0
(b) Run the simplex algorithm (with Bland’s Rule) to simultaneously obtain
optimal solutions to the primal and the dual.

(5) The 2-Stage Simplex Method. Use the 2-Stage Simplex Method to solve the
following LP (again, use Bland’s Rule for the pivots).
max 3x1 + x2
s.t. x1 − x2 ≤ −1
−x1 − x2 ≤ −3
2x1 + x2 ≤ 4
x1 , x2 ≥ 0

(6) Cones. Take a collection of m vectors in Rn , say V = {v1 , v2 , . . . , vm }. We


say the vector z ∈ Rn is in the cone of V if and only if there exist non-negative
numbers λ1 , λ2 , . . . , λm such that z = m
P
i=1 λi vi . Prove that z is not in the
cone of V if and only if there exists a vector w ∈ Rn such that z is less than
90 degrees from w but every vi ∈ V is at least 90 degrees from w.

You might also like