(2) (1 pt.) Let x1 , . . . , xm Rn . Derive from Farkas Lemma that exactly one of the following holds: (a) 0 conv.hull{x1 , . . . , xm }; (b) there is a row vector y Rn such that yxi > 0 for i = 1, . . . , m. Here 0 denotes the all-zero vector. (3) (1 pt.) Give the definition of a vertex of a polyhedron. Determine all the vertices of {x Rn | 1 xi 1, i = 1, . . . , n}. Prove correctness of your answer. You may use any theorems you know as long as you formulate them correctly. (4) (1 pt.) Consider the dual pair of optimization problems max{cx | Ax b, x Rn } = min{yb | yA = c, y 0, y Rm }. Formulate the complementary slackness condition for this dual pair. Show that a pair x, y of feasible (primal & dual) solutions is a pair of optimal solutions if and only if it satisfies the complementary slackness condition. (5) (1.5 pts.) Consider the linear optimization problem
1 1 1 1 max{ 6 15 8 x | 1 2 1 x 7 , x 0}. 1 2 1 4 (a) Solve this problem using the simplex method. (b) Determine the dual of this problem. (c) Find a dual optimal solution. (6) (1 pt.) Solve the following problem using the simplex method. Clearly describe your computation. Give all intermediate basic solutions and their objective values.
3 7 2 12 max{ 2 2 1 x | x , x 0}. 1 3 1 5 (7) (1 pt.) Compute v, W, k such that
4 5 3 7 0 5 {x Z | x= } = {v + W y | y Zk }. 2 4 6 5 0 (8) (1 pt.) Each of n teams plays against every other team a total of k games. Assume that every game ends in a win or a loss and let xi be the number of wins of team i. Let X be the set of all possible outcome vectors (x1 , . . . , xn ). Given an arbitrary vector (x1 , . . . , xn ), we would like to determine whether it belongs to X. Provide a network flow formulation of this problem. (9) (1.5 pts.) Consider the integer linear program max{x1 + 3x2 | 5x1 3x2 7, 2x1 + 3x2 11, x1 + 3x2 5, x 0, x Z2 }. (a) Draw the feasible set and point out the optimal solution. (b) Solve the problem by branch and bound; solve the linear programming problems that arise graphically. (c) Suppose you dualize the constraint 5x1 3x2 7. Formulate the Lagrangean relaxation that arises. Compute its value. 1