Class 15, C&O 367



  1. Optimality Conditions for Convex Programs cont...
    1. Review of Lemma: K is a ccc iff K=K++ and proof using hyperplane separation theorem
    2. Review of Farkas' Lemma and proof using the above Lemma
    3. KKT conditions for a convex program (sufficient, necessary with Slater's constraint qualification) with proof using relation between tangent and linearizing cones (see also Theorems 5.2.13 and 5.2.14)
    4. examples where KKT holds/fails
    5. Accessibility Lemma, Supporting Hyperplane Theorem 5.1.9. (Theorem 5.1.10 but without proof)