Generalized conjugate residual method for solving large and dense linear systems in Semidefinite programming

摘要

It is a great concern in the field of Semidefinite Programming to solve the large and dense linear systems which arise from the interior-point method. Often direct methods are too expensive in terms of computer memory and CPU-time requirements, then the only alternative is to use iterative methods. Here, we apply the generalized conjugate residual method to this type of SDP's linear systems. The theoretical properties and practical performance of the method will be discussed.