This preview shows pages 1–3. Sign up to view the full content.
This preview has intentionally blurred sections. Sign up to view the full version.
View Full Document
Unformatted text preview: C&O 350 Linear Optimization – Winter 2009 – Solutions to Assignment 7 Question # Max. marks Part marks 1 9 2 4 3 2 10 2 6 2 3 5 4 16 5 No Credit Total 40 1. Revised Simplex Method, I 9 marks = 2+4+3 Consider the linear programming problem ( P ) : maximize c T x subject to Ax = b, x ≥ where A =  5 1 4 4 2 1 0 3 2 1 1 0 0 1 1 1 1 1 , b = 2 1 1 , and c = [1 , 2 , 3 , 1 , , 2 , 1] T . (a) Determine the basic feasible solution associated with the basis B = { 1 , 6 , 7 } . (b) Solve ( P ) with the Revised Simplex Method, starting from the (feasible) basis B . In every iteration, use the smallest subscript rule for choosing the entering and leaving variables. Clearly show the details of each iteration. (c) If you find a feasible solution that is optimal, write down the optimal solution that you find. Give a proof of its optimality using duality theory. If you find that ( P ) is unbounded, write down a sequence of feasible solutions { x ( t ) } which shows this unboundedness, and show that the corresponding objective values become arbitrarily large as t goes to infinity. Solution : (a) A B =  5 1 0 3 0 0 1 1 1 . Solve A B x B = b to get x * B = 1 / 3 11 / 3 5 . Thus, the basic feasible solution associated with B = { 1 , 6 , 7 } is x = [1 / 3 , , , , , 11 / 3 , 5] T . (b) Iteration 1: Solve A T B y = c B , that is  5 3 1 1 0 1 0 0 1 y = 1 2 1 to get y = 1 5 / 3 1 . 1 Compute c 2 = c 2 A T 2 y = 2 [ 1 0 1 ] y = 2 > 0. x 2 enters. Solve A B d = A 2 to get d =  1 2 . d ≤ 0 = ⇒ LP problem is unbounded. (c) Proof of unboundedness: For each real number t , let x 2 ( t ) = t , x 3 ( t ) = x 4 ( t ) = x 5 ( t ) = 0 and x 1 ( t ) x 6 ( t ) x 7 ( t ) = 1 / 3 11 / 3 + t 5 + 2 t . It is easy to check that x ( t ) = [ x 1 ( t ) ,...,x 7 ( t )] T is a feasible solution for all t ≥ 0 (that is, x i ( t ) ≥ and Ax ( t ) = b ). Objective value of x ( t ) = 1 / 3 + 2 t + 2(11 / 3 + t ) (5 + 2 t ) = 8 / 3 + 2 t → + ∞ as t → + ∞ . Hence the LP problem is unbounded. 2. Revised Simplex Method, II 10 marks = 2+6+2 For part (b): 4 marks for the first iteration, 2 marks for the remainder Consider the linear programming problem ( P ) : maximize c T x subject to Ax = b, x ≥ where A = 1 2 0 1 2 1 0 2 1 1 2 1 0 0 , b = 1 2 5 , and c = [2 , , 1 , 4 , 2] T ....
View
Full
Document
This note was uploaded on 02/05/2010 for the course CO 350 taught by Professor S.furino,b.guenin during the Spring '07 term at Waterloo.
 Spring '07
 S.Furino,B.Guenin

Click to edit the document details