Application of PSO in the improved real-time evolution

An improved real-time evolution (RTE) strategy is proposed in an effort to apply RTE more accurately in optimization. The improved RTE tries to consider parameter update and employs a new algorithm as a substitution of the direct search. Algorithm performance comparisons are made among direct search, particle swarm optimization (PSO) and genetic algorithm (GA) to derive the suitable algorithm. PSO is eventually proposed due to its better performances of avoiding local optima to a certain extent, being apt to implement parameter update as well as fewer generations than GA. Under the improved RTE strategy, process is divided into several pseudo-steady state processes with time partitions. Within every pseudo-steady state process, PSO is employed to find the optimum and update parameters. The implement procedures are presented in detail. To demonstrate the feasibility and validity of PSO algorithm in the improved RTE, a classical benchmark is employed as a case study.