Abstract:Generalized gradient projection neural network models are proposed to solve nonsmooth convex and nonconvex nonlinear programming problems over a closed convex subset of ℝ n . By using Clarke’s generalized gradient, the neural network modeled by a differential inclusion is developed, and its dynamical behavior and optimization capabilities both for convex and nonconvex problems are rigorously analyzed in the framework of nonsmooth analysis and the differential inclusion theory. First for nonconvex optimization problems, the quasiconvergence results similar to those achieved by previous work are proved. For convex optimization problems, the global convergence results are proved, i.e. any trajectory of the neural network converges to an asymptotically stable equilibrium point, which is an optimal solution of the primal problem whenever it exists. This result remains valid even if the initial condition is chosen outside the feasible set. In addition, an asymptotic control result involving a Tykhonov-like regularization shows that any trajectory of the revised neural network can be forced to converge toward a particular optimal solution of the primal problem. Finally, two simulation optimization algorithms for solving the optimization and control problems are designed respectively, and three typical simulation experiments are illustrated to present the accuracy and efficiency of theoretical convergence results of this paper.

Gradient Projection Method for Optimization Problems with a Constraint in the Form of the Intersection of a Smooth Surface and a Convex Closed Set

Random Multi-Constraint Projection: Stochastic Gradient Methods for Convex Optimization with Many Constraints

A Proximal-Gradient Method for Constrained Optimization

A feasible smoothing accelerated projected gradient method for nonsmooth convex optimization

O(log T) Projections for Stochastic Optimization of Smooth and Strongly Convex Functions

An interior proximal gradient method for nonconvex optimization

Projection-Free Non-Smooth Convex Programming

Subgradient Projection Algorithms for Convex Feasibility on Riemannian Manifolds with Lower Bounded Curvatures

An Efficient Gradient Projection Method for Structural Topology Optimization

An Infeasible-Point Subgradient Method Using Adaptive Approximate Projections

Projected Subgradient Minimization versus Superiorization

Convergence Analysis of a Projected Gradient Method for Multiobjective Optimization Problems

Nonsmooth Projection-Free Optimization with Functional Constraints

Single-Projection Procedure for Infinite Dimensional Convex Optimization Problems

Generalized Gradient Projection Neural Networks for Nonsmooth Optimization Problems

Linear Optimization by Conical Projection

Efficient Projected Gradient Methods for L 0 Constrained Optimization

Convex Optimization with an Interpolation-based Projection and its Application to Deep Learning

The Gradient Projection Method for a Supporting Function on the Unit Sphere and Its Applications