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 Document
Unformatted text preview: Lecture 23 NonSimplex Steps UCSD Math 171A: Numerical Optimization Philip E. Gill http://ccom.ucsd.edu/~peg/math171 Monday, March 9th, 2009 Recap: Starting at a nonvertex A simple modification of the simplex method allows us to: 1 Start at a feasible point that is not necessarily a vertex. 2 Solve problems that may not have a vertex solution. The modification is to add temporary bound constraints to define a “fake” vertex. At an arbitrary (feasible) x we add temporary bound constraints: x i ≥ ≤ ( x ) i i.e., e T i x ≥ ≤ ( x ) i UCSD Center for Computational Mathematics Slide 2/44, Monday, March 9th, 2009 Example 1: minimize 2 x 1 + x 2 subject to the constraints: constraint #1: x 1 + x 2 ≥ 1 constraint #2: x 2 ≥ constraint #3: x 1 ≥ constraint #4: x 1 ≥  2 constraint #5: x 1 + x 2 ≥  2 In allinequality form, we have min c T x subject to Ax ≥ b , where c = 2 1 ! , A = 1 1 1 1 1 1 1 , b = 1 2 2 UCSD Center for Computational Mathematics Slide 3/44, Monday, March 9th, 2009 x 1 x 2 x * #3 #1 #2 #5 #4 Apply the simplex method for allinequality form. Start at the interior point x = 3 2 1 2 ! . At x , we have A ( x ) = ∅ , so no constraints are active. Add the temporary bounds: constraint #6: x 1 ≥ ≤ 3 2 constraint #7: x 2 ≥ ≤ 1 2 If W = { 6 , 7 } , then A = 1 1 ! , b = 3 2 1 2 ! UCSD Center for Computational Mathematics Slide 5/44, Monday, March 9th, 2009 x 1 x 2 x * #3 #1 #2 #5 #4 x First iteration Step 1: Check for optimality A T λ = c ⇒ 1 0 0 1 ! λ = 2 1 ! ⇒ λ = 2 1 ! ← s = 1 W = { 6 ↑ ws =6 , 7 } ( λ ) 1 = 2 is associated with temporary constraint x 1 ≥ ≤ 3 2 . UCSD Center for Computational Mathematics Slide 7/44, Monday, March 9th, 2009 First iteration Step 2: Compute the search direction Compute the search direction A p = ± e 1 ( λ ) s > 0, so we choose e 1 for the righthand side. A p = e 1 ⇒ 1 0 0 1 ! p = 1 ! ⇒ p = 1 ! UCSD Center for Computational Mathematics Slide 8/44, Monday, March 9th, 2009 x 1 x 2 x * #3 #1 #2 #5 #4 x p First iteration Step 3: Take a step Find the constraint residuals: r = Ax b = 1 1 1 1 1 1 1 3 2 1 2 ! 1 2 2 = 1 1 2 3 2 1 2 1 UCSD Center for Computational Mathematics Slide 10/44, Monday, March 9th, 2009 First iteration Step 3: Take a step Take a step....
View
Full Document
 Winter '08
 staff
 Math, Linear Programming, Optimization, UCSD Center for Computational Mathematics, Computational Mathematics

Click to edit the document details