基于α约束支配排序混合进化算法的微电网多目标优化运行  

Microgrid operation multi-objective optimization based on hybrid evolution algorithm with α-constraint dominant sorting

作　　者：彭春华[1] 黄戡[1] 袁义生[1] 潘蕾[2] 

机构地区：[1]华东交通大学电气与电子工程学院,江西南昌330013 [2]东南大学能源热转换及其过程测控教育部重点实验室,江苏南京210096

出　　处：《电力自动化设备》2015年第4期24-30,37,共8页Electric Power Automation Equipment

基　　金：国家自然科学基金资助项目(51167005;51106024);教育部人文社科青年基金资助项目(14YJCZH135);江西省科技支撑计划项目(20142BBE50001);江西省教育厅科技基金资助项目(GJJ14386)~~

摘　　要：为降低微电网运行成本和污染排放,建立微电网多目标优化运行模型,并提出一种新型的α约束支配排序混合进化算法求解模型,该算法通过采用α约束支配排序机制,将所有约束条件统一处理为α约束水平度,并将其作为进化选择指标以控制所有个体快速转化为可行解,可显著提高约束处理效率。提出一种基于非劣排序的混合多目标进化算法,有效融合微分进化算法与分布估计算法各自的优点,克服单一算法种群多样性不足和易早熟的缺陷。通过分类逼近理想解的排序方法实现多属性决策,以获得最优折中解。某微电网算例结果表明所提算法有效、可行。In order to reduce the operational cost and pollution emission,a multi-objective optimization model is built for microgrid and a hybrid evolution algorithm with α-constraint dominant sorting is proposed to solve the model, which applies the α- constraint dominant sorting mechanism to treat all constraints as theα- constraint levelness and takes the levelness as the evolutionary selection index to quickly transform all individuals into the feasible solution,significantly improving the constraint processing efficiency. A hybrid multi-objective evolution algorithm with non-dominated sorting is proposed to effectively combine the advantages of the DEA（Differential Evolution Algorithm） and EDA（Estimation of Distribution Algorithm） for overcoming the defects of low species diversity and premature convergence of single algorithm. The similarity sorting method is adopted to approach the ideal solution for realizing the multi-attribute decision and obtaining the optimal compromise solution. Case study for a microgrid shows that the proposed algorithm is effective and feasible.

关 键 词：微电网 优化 α约束支配 进化算法 多属性决策 模型 

分 类 号：TM732[电气工程—电力系统及自动化]