Approximation of convexly constrained pseudoinverse by hybrid steepest descent method

Recently, a convexly constrained pseudoinverse problem unifying a wide class of signal processing problems was considered by Sabharwal and Potter (1998). In this paper, we present a new algorithmic solution to this extremely important inverse problem. The proposed algorithm is straightforwardly derived from a theorem of the Hybrid Steepest Descent Method (HSDM) recently developed by the authors, where the strong convergence of the algorithm to the unique solution of the problem is rigorously guaranteed