IMPROVED RELAXED CQ METHOD WITH INERTIAL ACCELERATED TECHNIQUE FOR THE SPLIT FEASIBILITY PROBLEM

In this paper, we study the split feasibility problem and present an improved relaxed CQ iterative algorithm with inertial accelerated technique in real Hilbert spaces. By mixing the self-adaptive primal-dual method with the inertial technique, not only the computation on the matrix norm and the orthogonal projection is relaxed but also the step size of the dual variable is improved from fixed selection to variable selection. Under suitable conditions, the weak convergence of the algorithm is established. Our result extends and improves the corresponding results.