Adaptive Critics for Dynamic Particle Swarm Optimization

This work introduces a novel technique for dynamic particle swarm optimization (DPSO) using adaptive critic designs. The adaptation between global and local search in an optimization algorithm is critical for optimization problems especially in a dynamically changing environment or process over time. The inertia weight in particle swarm optimization (PSO) is dynamically adjusted in this paper in order to provide a nonlinear search capability for the PSO algorithm. Results on benchmark functions in the literature are provided