Group for Research in Decision Analysis


A Primal-Dual Regularized Interior-Point Method for Semidefinite Programming

, , and

Interior-point methods in semi-definite programming (SDP) require the solution of a sequence of linear systems which are used to derive the search directions. Safeguards are typically required in order to handle rank-deficient Jacobians and free variables. We propose a primal-dual regularization to the original SDP and show that it is possible to recover an optimal solution of the original SDP via inaccurate solves of a sequence of regularized SDPs for both the NT and dual HKM directions. This work is a generalization of recent work by Friedlander and Orban for quadratic programming.

, 29 pages