This preview shows pages 1–7. Sign up to view the full content.
This preview has intentionally blurred sections. Sign up to view the full version.
View Full DocumentThis preview has intentionally blurred sections. Sign up to view the full version.
View Full DocumentThis preview has intentionally blurred sections. Sign up to view the full version.
View Full Document
Unformatted text preview: CO372 Assignment 2, Winter 2010 Solutions 1 2 Chapter 1 Optimization Exercise 1.10 Let C be a ( n,n ) positive definite matrix and let A be a ( m,n ) matrix having full row rank. Define x = (1 , 1 ,..., 1) , u = (1 , 1 ,..., 1) and b = Ax . (a) Show how to choose c so that x is optimal for (1.4) with multiplier vector u . (b) Computationally verify Part (a) by modifying Example1p2.m (Figure 1.4) to solve (1.4) with C = diag (1 , 2 , 3 , 4 , 5) , A = 1 2 3 4 5 6 7 8 9 10 . Solution: (a) C , A , x , b = Ax and u are all specified. The second optimality condition of Theorem 1.2 is satisfied provided c Cx = A u , or, c = Cx A u. 3 CHAPTER 1. OPTIMIZATION (b) The modification of Example1p2.m (Figure 1.4, text) to solve this prob lem is named Exer1p10.m and is shown in Figure 1.1. 1 %Exer1p10.m Exercise 1.10(b) 2 n = 5; m = 2; 3 x0 = [ 1 1 1 1 1 ]' 4 u = [ 1 1 ]' 5 BigC = [1 0 0 0 0; 0 2 0 0 0 ; 0 0 3 0 0; 0 0 0 4 0; ... 6 0 0 0 0 5 ] 7 A = [1 2 3 4 5 ; 6 7 8 9 10 ] 8 b = A * x0 9 c = BigC * x0 A' * u 10 11 checkdata(BigC,1.e 6); 12 H = vertcat([BigC,A'] , [A,zeros(m,m)]) 13 rhs = vertcat( c,b) 14 y = H 1 * rhs % Solve Hy = rhs. 15 x = y(1:n) % x = optimal solution 16 u = y(n+1:n+m) % u = multipliers for Ax = b Figure 1.1 Exercise 1.10(b): Exer1p10.m. The output from running Exer1p10.m gives c = ( 8 , 11 , 14 , 17 , 20) and b = (15 , 40) . Solving the linear equations of Theorem 1.3 (lines 12 to 16 of Figure 1.1) with data A , C , c and b produces the output x = (1 , 1 , 1 , 1 , 1 , ) and u = (1 , 1) as expected. 4 Chapter 2 The Efficient Frontier Exercise 2.1 Show 2 = 0 if and only if is a multiple of l . Solution: Recall that 2 = h 1 h 1 and h 1 =  1  l  1 l  1 l  1 l. Suppose first ( ) 2 = h 1 h 1 = 0. Since is positive definite, we have h 1 = 0. Therefore,  1 = l  1 l  1 l  1 l, which implies that = l  1 l  1 l l, so that is indeed a multiple of l . Next suppose ( ) = l for some constant . Then h 1 =  1 l l  1 l l  1 l  1 l 5 CHAPTER 2. THE EFFICIENT FRONTIER =  1 l  1 l = 0 , which implies that 2 = h 1 h 1 = 0. / Exercise 2.2 Verify equation (2.15) ( 1 = 0)....
View
Full
Document
 Spring '10
 MICHAEL

Click to edit the document details