Dec 6, 2012 Student Name: MATH 4/5420 - Optimization Final Exam - Fall 2012 Take Home - Due Thursday, Dec 13, before 10am (no Exceptions!) Problem 1 Solve the following linear programming problem using the simplex tableau method max 3x1 + 2x2 subject to x1 ≤ 4, x2 ≤ 1 2x1 + x2 ≤ 5 x1 , x2 ≥ 0 Problem 2 For the linear programming problem min x1 + x2 + 3x3 subject to x1 − x2 = 1 x1 + x3 ≥ 2 x1 ≥ 0, x2 ≥ 0, x3 ≥ 0 (a) Find the dual problem and solve it using the graphical method. (b) Write down the KKT optimality conditions for the primal problem and solve it. Problem 3 Consider the variational problem of minimizing Z J[x] = 1p 1 + [x0 (t)]2 dt 0 over all functions x = x(t), 0 ≤ t ≤ 1, with x(0) = 1 and x(1) = 0. (a) Use the Euler-Lagrange equation to find the optimal solution. Can you verify that it is indeed a minimizer? (b) Using a discretization of the objective in (a) using N + 1 equal subintervals of [0, 1], solve the resulting nonlinear optimization problem for X = (xj )1≤j≤N (in N-dimensions) to find an approximate solution to (a). Does this approximation converge to the exact solution as N → ∞? (c) [MATH 5420 Only] Redo (a) and (b) if, in addition, we imposes the additional constraint on x Z 1 x(t) dt = 1 0 Problem 5 The fish population in a lake x = x(t) needs to be reduced in the next five years from the current value x(0) = 1000 to half of that, x(5) = 500. The population of fish tends to grow by about 10% a year, but it can be changed by controlling the number u(t) of fishermen. Then the population of fish will evolve according to dx = 0.1x − u dt Since human intervention has its drawbacks for the environment, one wants to minimize the Z 5 u2 min J[u] = 0 (a) Find the optimal function u∗ and the resulting fish population x∗ . (b) Solve the same problem by converting it first to a calculus of variation problem [eliminate u] and using the Euler-Lagrange equation. Can you verify that this is indeed a minimizer?