Khintchine-Type Inequalities and Their Applications in Optimization

By

Professor Man-Cho So, Anthony

Department of Systems Engineering & Engineering Management
The Chinese University of Hong Kong

Date: Sept 25, 2008

Time: 4:30pm - 5:30 pm

Venue: Rm. 121, Ho Sin Hang Engineering Building, CUHK

Abstract :

In this talk we first give a brief introduction to Khintchine-type inequalities, which are inequalities that bound the moments of the norm of a certain sum of independent Banach space-valued random variables. Then, we will discuss how these inequalities can be used to obtain the best known bounds for several optimization problems, including the approximation of a class of quadratic optimization problems with orthogonality constraints, the design of so-called safe tractable approximations for chance constrained linear matrix inequality systems, and the design of efficient sampling procedures for certain stochastic optimization problems.