差分进化算法在电力系统环境经济调度中的应用  

Application of differential evolution algorithm to environmental/economic dispatch of power systems

作　　者：卢有麟[1] 周建中[1] 覃晖[1] 许可[1] 

机构地区：[1]华中科技大学水电与数字化工程学院,湖北武汉430074

出　　处：《华中科技大学学报（自然科学版）》2010年第8期121-124,共4页Journal of Huazhong University of Science and Technology(Natural Science Edition)

基　　金：国家科技支撑计划重大项目(2008BAB29B08);科技部水利部公益性行业专项科研基金资助项目(200701008);国家重点基础研究计划资助项目(2007CB714107)

摘　　要：以求解环境经济调度(EED)这一复杂的多目标约束优化问题为背景,研究了一种改进的多目标差分进化算法(EMODE),该算法依据多目标优化问题的特点重新设计了差分进化算法(DE)的进化算子并引入自适应二次变异算子来有效避免DE存在的'早熟'收敛现象;同时,针对EED问题约束条件复杂且难以处理这一问题,依据不同类型约束的特点提出一种启发式的约束处理方法.将EMODE应用到某电力系统的多目标环境经济调度中,仿真计算结果以及与其他求解方法的对比分析表明,EMODE可以有效兼顾全局收敛性和Pareto非劣调度方案的多样性,具有较高的效率以及鲁棒性.Environmental/economic dispatch （EED） which is a complicated multi-objective constrained problem. An enhanced multi-objective differential evolution algorithm （EMODE） is proposed to solve that problem. The proposed algorithm redesigns the evolutionary operator of differential evlution （DE） and adopts a second mutation operator to mutate individuals during the computing process to avoid premature convergence. Meanwhile, in view of the difficulties of handling the complicated constraints of EED problem, a new constraints handle method is presented. The proposed algorithm is applied to solve the EED problem of a power system, experimental results demonstrate the successful application of EMODE, compared with other methods, EMODE outperforms in globe convergence and the diversity of the Pareto set with higher effectiveness and robustness.

关 键 词：电力系统 多目标调度 差分进化 多目标进化算法 污染排放 全局最优 约束处理 

分 类 号：TM715[电气工程—电力系统及自动化]