This preview shows pages 1–2. Sign up to view the full content.
This preview has intentionally blurred sections. Sign up to view the full version.View Full Document
Unformatted text preview: CS 435 : Linear Optimization Fall 2008 Lecture 9: Duality Theorem Lecturer: Sundar Vishwanathan Computer Science & Engineering Indian Institute of Technology, Bombay Here is the duality theorem. 1 Duality Theorem 1 If the primal is feasible and has a nite optimum then the dual is also feasible, has a nite optimum and their optimums must be equal. Proof: We rst show that the dual is feasible. This means we can nd a y which satis es A T y = c ; y . What does A T y = c mean? It means we can write the cost function as a linear combination of the rows of A . Rows of A are also the outward normals to the hyperplanes in Ax b . The inequality y means the coe cients should be non-negative. We showed that at an optimum point of the primal, we can write the cost vector as a non-negative linear combination of the normals to the de ning hyperplanes. The coe cients of this non-negative linear combination yield a feasible point in the dual....
View Full Document
- Summer '09
- Computer Science