1 / 12

CS5 3 21 Numerical Optimization

CS5 3 21 Numerical Optimization. 14 Linear Programming: The Interior Point Method. Interior point methods for LP. There are many variations of IPM for LP Primal IPM, dual IPM, primal-dual IPM Potential Reduction Methods Path Following Methods Short-step methods Long-step methods

pruneda
Download Presentation

CS5 3 21 Numerical Optimization

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript


  1. CS5321 Numerical Optimization 14 Linear Programming: The Interior Point Method

  2. Interior point methods for LP • There are many variations of IPM for LP • Primal IPM, dual IPM, primal-dual IPM • Potential Reduction Methods • Path Following Methods • Short-step methods • Long-step methods • Predictor-corrector methods • We will cover the path following method and the long-step method

  3. The dual problem of LP minxz=cTx subject to Ax = b, x  0 • The dual problem is (see chap 12) maxbT subject to AT +s = c, s  0 • Weak duality For any x and  that are feasible in the primal and the dual problems, bT  cTx. • Strong duality If x* and * are solutions to the primal and the dual problems, bT* = cTx*.

  4. T ¸ A + s c = b A x = 0 ¸ x s ; 0 x s = i i Optimality conditions of LP minxz=cTx subject to Ax = b, x  0 • The Lagrangian function L(x,,s) = cTx − T(Ax−b) − sTx • The optimality conditions (KKT, see chap 12) • The last one is the complementarity condition

  5. 0 1 ¢ x 0 T 1 ¸ A + ¡ s c ( ) ( ) ¸ ¸ ¢ ¸ J F ¡ x s x s = @ A ; ; ; ; ( ) b ¸ f A F 0 0 ¸ ¡ x x s o r x s = = @ A ¢ ; ; ; ; s X S e Primal-dual method • The optimality conditions can be expressed as • X=diag(x), S=diag(s) and e=(1,1,…,1)T. • F is a nonlinear equation: R2n+m  R2n+m. • Solved by the Newton’s method: (J the Jacobian of F.) (x, , s)+=(x, , s),

  6. 0 0 T 1 1 0 0 T 1 1 ¸ r r r A A I 0 + ¡ r r r r s c ¸ x c c s c c ( ) b ¸ r r A r A J F 0 0 ¡ r r x r r x s = = = = b ¸ b b b @ @ A A @ @ A A x s ; ; r r X r S S X 0 r r r e r ¸ X S X S X S X S x s Jacobian of F • Let . • The Jacobian J (x, , s) of F(x, , s) is

  7. µ 0 T T ¶ µ 1 0 ¶ 1 µ 0 1 ¶ 2 1 ¡ ¡ 1 1 ¡ ¡ D A A I ¢ ¢ X 0 ¢ X ¡ X S ¢ ¡ + ¡ ¡ ¡ x x r r r X S s r x = X S = c = c 1 2 1 2 ¡ D S X = = ¸ ¸ A A ¢ ¢ 0 0 0 ¡ ¡ r r = b b @ A @ A @ A S X ¢ 0 ¡ s r X S Solving the linear systems • The Newton’s direction is obtained by solving • S is an artificial variable. Let • The left hand side is symmetric, but indefinite.

  8. n T 1 x s X ¹ x s = = i i n n i 1 = Without centering Centered direction Centering • In practice, rXS=XSe −e rather than rXS=XSe. • Duality measure: • Centering parameter   [0,1]: • Decreasing as approaching solutions • This is related to the barrier method (chap 15)

  9. The path following algorithm • Given (x0,0,s0) with (x0,s0)>0 • For k = 0, 1, 2 …until (xk)Tsk<  • Choose k[0,1] and let k=(xk)Tsk/n. • Solve the Newton’s direction (xk, k, sk) • Set (xk+1, k+1, sk+1)=(xk, k, sk) + k(xk, k, sk), where k is chosen to make (xk+1,sk+1)>0

  10. Central path and neighborhood • Feasible set F0={(x,,s) | Ax=b, AT+s=c, (x,s)>0} • Point (x,,s)F0 is on the central pathC if x(i)s(i)=  for all i=1,2,…,n. • An example of the neighborhood of the central path •  is the duality measure (two slides before) • Typically =10−3.

  11. Long-step path-following • Given (x0,0,s0)N− and , min, max. • For k = 0, 1, 2 …until (xk)Tsk<  • Choose k[min,max] and let k=(xk)Tsk/n. • Solve the Newton’s direction (xk, k, sk) • Set (xk+1, k+1, sk+1)=(xk, k, sk) + k(xk, k, sk), where k is chosen to make (xk+1, k+1,sk+1)N−.

  12. Complexity • The long step following method can converge in O(nlog(1/)) iterations (Theorem 14.4) • The most effective interior point method for LP is the predictor-corrector algorithm (Mehrotra 1992)

More Related