Error Bounds for Conic Feasibility Problems - Case Studies on the Exponential Cone

Explore error bounds for conic feasibility problems in this 53-minute distinguished seminar by Ting-Kei Pong from Hong Kong Polytechnic University. Delve into a general framework for deriving error bounds, focusing on the exponential cone and its applications in optimization problems involving power, exponential, logarithmic, and entropy functions. Learn about facial reduction, one-step facial residual functions, and their role in handling degeneracy in conic programs. Discover how these concepts apply to the study of Kurdyka-Lojasiewicz property and gain insights into the design of termination criteria for conic solvers and convergence rate analysis of algorithms.