求解机组组合问题的变邻域-粒子群混合算法  

A hybrid algorithm of VNS-PSO to solve unit commitment problem

作　　者：季美红[1] 徐宁舟[1] 高研[1] 丁银[1] 

机构地区：[1]合肥工业大学电气与自动化工程学院,安徽合肥230009

出　　处：《合肥工业大学学报（自然科学版）》2010年第6期827-831,共5页Journal of Hefei University of Technology：Natural Science

摘　　要：文章提出了一种变邻域-粒子群搜索算法,用以解决多种约束条件下的机组组合问题;定义了3种邻域结构来处理机组启停状态,采用变邻域策略扩大搜索空间,避免了搜索停滞现象,并根据邻域结构确定合适的候选解集,确保了解的质量。在确定机组启停状态后,再采用粒子群算法进行机组的功率分配,针对PSO易陷入局部极值的缺点,采用极值扰动的策略进行了改进,从而帮助粒子摆脱局部极值,获得更优解,结果表明该了方法的可行性和有效性。This paper proposes a variable neighborhood search （VNS）-particle swarm optimization （PSO） algorithm to solve the problem of unit commitment with many kinds of restrictions. The algo- rithm defines three neighborhood structures to deal with the running or suspended status of units, ap- plies variable neighborhood search strategy of expanding space to avoid searching stagnation, and de- termines the appropriate candidate solution set according to the neighborhood structure to ensure the quality of solution. After confirming the running or suspended status of units, the particle swarm al- gorithm is used to distribute power among units. Aiming at the local extreme value resulted from PSO, an extreme disturbance strategy is applied to help particles effectively break away from the local extreme value and obtain the optimal solution. The results of examples prove that the method is feasi- ble and effective.

关 键 词：机组组合 负荷分配 变邻域 粒子群算法 

分 类 号：TM73[电气工程—电力系统及自动化]