求解环境经济调度问题的多目标差分粒子群优化算法

Multiobjective Particle Swarm Optimization Based on Differential Evolution for Environmental/Economic Dispatch Problem

提出一种基于差分演化的改进多目标粒子群优化算法来求解电力系统环境经济调度问题。算法通过对Pareto最优解集的差分演化来增加Pareto最优解的多样性;通过循环拥挤距离来控制归档集中非劣解的分布,以提高对种群空间的均匀采样;采用一种新的多目标适应值轮盘赌法选择粒子的全局最优位置,使其更逼近Pareto最优前沿;自适应惯性权重和加速度因子的动态变化可增强算法的全局搜索能力。对电力系统环境经济负荷分配模型进行仿真,并与文献中的其他算法进行了比较。结果表明,改进的算法能够在保持Pareto最优解多样性的同时具有较好的收敛性能。

An improved multiobjective particle swarm optimization based on differential evolution technique is proposed for environmental/economic dispatch（EED） problem.The algorithm adopts differential evolution to increase the diversity of the Pareto set.Circular crowded sorting approach helps to generate a set of well-distributed Pareto-optimal solutions in one run.The global best individuals in multiobjective optimization domain are redefined through a new multiobjective fitness roulette technique.And the adaptive inertia weight and acceleration coefficients enhance the global exploratory capability.The environmental/economic loading distribution model in power system is simulated and compared with other algorithms in references.The results indicate that the improved algorithm can maintain the diversity of Pareto-optimal solutions and is of better convergency at the same time.