Multi-objective particle swarm optimization control technology and its application in batch processes

Abstract

In this paper, considering the multi-objective problems in batch processes, an improved multi-objective particle swarm optimization based on pareto-optimal solutions is proposed. In this method, a novel diversity preservation strategy that combines the information on distance and angle into similarity judgment is employed to select global best and thus guarantees the convergence and the diversity characteristics of the pareto front. As a result, enough pareto solutions are distributed evenly in the pareto front. Lastly, the algorithm is applied to a classical batch process. The results show that the quality at the end of each batch can approximate the desire value sufficiently and the input trajectory converges; thus verify the efficiency and practicability of the algorithm. © 2010 Springer-Verlag.