Instructor: Emre Mengi Office: SCI 113 E-mail: emengi@ku.edu.tr Phone: (212) 338-1658 Office Hours: Fri 16:30 - 17:20 (or by appointment) (office hours will also be held remotely through zoom) | Course Description: A
major way of categorizing optimization problems is through their
convexity. Finding globally optimal solutions of convex problems are
tractable, whereas this is often deemed intractable for non-convex
problems. Here, we focus on the former. After introducing the basic
concepts regarding convex sets and convex functions, we present the
optimality conditions and duality theory for convex optimization
problems, where a special attention is paid to common convex
optimization problems such as linear programs, convex quadratic
programs, semidefinite programs. The second part is mainly on the applications of convex optimization and the algorithms for their solution. The Newton method based algorithms and interior-point methods described in the second part exploit the optimality conditions and the duality theory developed in the first part. Announcements:
|