多目标混合进化算法及其在经济调度中的应用

Multiobjective Hybrid Evolutionary Algorithm for Economic Load Dispatch

鉴于环境保护的要求,对于经济调度问题,需同时考虑环境要求、发电费用等多个目标。提出一种基于进化规划(evolutionary programming,EP)和粒子群优(particle swarm optimization,PSO)的多目标混合进化算法(multi-objective evolutionary programming and particle swarm optimization,MOEPPSO),MOEPPSO采用了EP的变异操作,用来抑制PSO的快速收敛所带来的种群早熟问题,而PSO的记忆、协作能力则弥补了EP收敛速度慢的缺点。此外,MOEPPSO应用自适应网格算法对外部库中的Pareto解集进行调整,对一个30节点IEEE系统进行计算,结果显示MOEPPSO在获得最优Pareto解集、降低计算复杂度、提高收敛效率等方面具有很强的优越性。

Due to the requirement of environment protection, while coping with the economic power dispatch problem, the competing and non-commensurable emission, fuel cost objectives should be considered simultaneously. A multi-objective hybrid evolutionary algorithm based on evolutionary programming （EP） and particle swarm optimization （PSO）, named MOEPPSO, is presented for this problem. The mutation operator in EP effectively restrains the prematurity phenomenon of PSO aroused by the fast convergence, while the memory and collaboration characteristics of PSO help EP to converge faster. Furthermore, the adaptive grid algorithm is applied to MOEPPSO. The effectiveness of the proposed algorithm is validated with the IEEE 30--bus system, and the results demonstrate the better Pareto front, the computation complexity reduction and the convergence efficiency improvement of the proposed algorithm.