A New Generalized Gradient Projection Method for General Constrained Optimization

In this paper,we propose a new generalizd gradient projection method for general constrained optimization based on a hybrid l1-l∞ penalty function.The characters of the proposed algorithm are as follows:the start point can be chosen arbitrarily;the searching direction is a decreasing direction of the merit function;and the algorithm is global convergent under the traditional assumptions.Finally,some preliminary numerical results show that the proposed algorithm is effective.