ALTERNATING PROJECTION METHOD FOR SOLVING DOUBLY STOCHASTIC INVERSE SINGULAR VALUE PROBLEMS WITH PRESCRIBED ENTRIES

Doubly stochastic inverse singular value problem with prescribed entries aims to construct a doubly stochastic matrix from the prescribed singular value and prescribed entries. In this paper, the doubly stochastic inverse singular value problem is considered as the problem of finding a point in the intersection of a compact set and a closed convex set. We present a numerical procedure which is based on an alternating projection process, for solving the problem. The method is iterative in nature. And each subproblem in the alternating projection method can be solved easily. Convergence properties of the algorithm are investigated and numerical results are presented to illustrate the effectiveness of our method.