| With the development of computer and the research need of practical problems,the problems of large-scale optimization have more attention.In recent years,many domestic and foreign scholars have given many new algorithms for solving linear constrained convex optimization.The classic algorithms include Projection Contraction Algorithm for solving monotone variational inequality,Proximal Point Algorithm in variational inequality framework.Alternating Direction Multiply Method for dealing for two variables.And ADMM-like splitting contraction algorithms for multi-block separable convex optimization problems.The relaxed PPA based contraction method,which is an efficient algorithm for solving the linear constrained convex optimization problem,but the parameter conditions of the relaxed PPA have a great influence on the calculation efficiency.Therefore,we consider to further relax the known best parameter conditions and find that the parameter conditions of the the Primal-Dual relaxed PPA algorithm are determined by(?)change to(?) Compared with the algorithm under the original parameter condition,the new algorithm obtained by further relaxing the step size is still convergent.Finally,the effectiveness of the algorithm is verified by two numerical experiments:correlation matrix correction and compression sensing. |
