A new steepest descent differential inclusion-based method for solving general nonsmooth convex optimization problems (Q2442700)

The authors consider general nonsmooth convex optimization problems. From differential equation-based and differential inclusion-based methods for solving optimization problems, there are four different classes: gradient projection methods, primal-dual methods, Lagrange multiplier rule methods and penalty-based methods. The method investigated in the present work is a steepest descent neural network for solving general nonsmooth convex optimization problems. The authors prove analytically the convergence to the optimal solution set and apply the method to some numerical tests, confirming the effectiveness and the performance of the proposed neural network.