Real-Time Optimal Control Using Particle Swarm Optimization

Abstract

Optimal Control Problems can be solved by a well known metaheuristic, Particle Swarm Optimization. A version of this metaheuristic has been used to develop an algorithm that solves such a problem. The solution can be used to control the considered dynamic system only in open loop. Therefore, the main problem is how to use the PSO algorithm in a control structure that works in closed loop. The solution of this problem is to use the well known control structure Model Predictive Control whose controller has an optimization algorithm that minimizes the prediction error. The proposal of this work is to confer this role to the PSO based algorithm because it would minimize implicitly the prediction error.