A collaborative neurodynamic approach to global and combinatorial optimization

Abstract

In this paper, a collaborative neurodynamic optimization approach is proposed for global and combinatorial optimization. First, a combinatorial optimization problem is reformulated as a global optimization problem. Second, a neurodynamic optimization model based on an augmented Lagrangian function is proposed and its states are proven to be asymptotically stable at a strict local minimum in the presence of nonconvexity in objective function or constraints. In addition, multiple neurodynamic optimization models are employed to search for global optimal solutions collaboratively and particle swarm optimization (PSO) is used to optimize their initial states. The proposed approach is shown to be globally convergent to global optimal solutions as substantiated for solving benchmark problems.