Two-direction Subgradient Method for non-differentiable Optimization Problems

Abstract

One more search direction is introduced into the subgradient method for non-differentiable optimization problems to enhance the speed of convergence. The proposed algorithm generates a point which is strictly closer and forms a more acute angle to the solution set than the point generated by the subgradient method. The resulting algorihtm still remains very simple. A computational test shows great improvement in the computational efficiency using this algorithm.