基于决策者偏好的微电网经济/环保多目标优化调度研究  

Study on multi-objective optimization for environmental/economic microgrid scheduling based on decision makers' preference

作　　者：李昕[1] 方彦军[1] 

机构地区：[1]武汉大学自动化系,湖北武汉430072

出　　处：《可再生能源》2016年第10期1423-1429,共7页Renewable Energy Resources

基　　金：国家自然科学基金项目(61374064);国家科技支撑计划项目(2013BAA01B01);中央高校经费资助项目(2014208020201)

摘　　要：文章针对微网多目标优化调度算法研究中的问题,提出基于偏好权重向量的第二代非劣支配排序遗传算法(PWV-NSGA-II),利用NSGA-II在多目标空间求解的特性以及决策者的偏好信息,一次性获取多个偏好权重下的Pareto折中解集。同时,设计了基于偏好程度的优秀个体选择策略,兼顾了决策者的偏好和种群的多样性。将PWV-NSGA-II应用在所建立的微网经济/环保多目标调度模型中,并详细分析了不同运行模式下算法的优化结果。分析结果表明,PWV-NSGA-II算法能有效地反映决策者的偏好信息,并能较好地适应不同运行模式下优化模型的变化,具有较强的鲁棒性,对于解决实际的微网多目标优化问题有一定的参考意义。Aiming at solving the problems in the current studies of multi-objective optimization algorithms for the microgrid scheduling, the non-dominated sorting genetic algorithm II is improved based on preference weight vector（PWV-NSGA-II）. By taking advantage of the search efficiency of NSGA-II and the preference information provided by the decision makers, the Pareto sets under various preference weights can be obtained in a single run. Moreover, a selection strategy of elitist individuals is proposed based on the preference degree, which considers both the decision makers＇ preference and the population diversity. Thereafter, PWV-NSGA-II is applied to the multi-objective optimization models of the microgrid scheduling designed in this paper, and the optimization results gained under different operation modes are presented and analyzed. The results show that PWV-NSGA-II can reflect the decision makers＇ preference information efficiently, and adapt to the model changes with different operation modes, which means the algorithm has strong robustness. Therefore,the proposed PWV-NSGA-II is efficient in dealing with the real-world multi-objective optimization problems for the microgrid scheduling.

关 键 词：微电网 风力发电 光伏发电 多目标优化 决策者偏好 

分 类 号：TK514[动力工程及工程热物理—热能工程] TM734[电气工程—电力系统及自动化]