摘要
针对电网规划需综合考虑经济性和可靠性的问题,提出一种改进的多目标蚁群算法。该算法采用改进的快速排序方法构造Pareto最优解集,缩短了"慢速链",降低了算法的时间复杂度;采用聚类算法裁剪非支配解,使所得解在整个Pareto解空间具有良好的多样性和分布性;采用信息素更新变参数控制,加快算法的全局收敛速度;采用挥发系数动态自适应调节机制,提高算法全局搜索能力。通过18节点电网规划算例证明,提出的改进算法与基本多目标蚁群算法相比,所得的Pareto最优解数量更多,Pareto前沿分布更加均匀,同时收敛性和快速性也得到了提高。
For the reason that both economy and reliability should be considered during power network planning, an improved multi-objective ant colony algorithm （IMACA） is proposed. In the proposed algorithm, the modified quick sort method is adopted to construct Pareto optimal solution set, thus the slow-chain is shortened and the time complexity of this algorithm is mitigated; the clustering algorithm is adopted to modify non-dominated solution, thus the obtained solution can possess good diversity and distributivity in whole Pareto solution space; the sociohormone is adopted to update variable parameter control, thus the global convergence is speeded up; the sociohormone volatilization coefficient is used to dynamic adaptive regulation mechanism, thus the global search ability of the proposed algorithm is improved. The calculation results of an 18-bus power network planning show that more Pareto optimal solutions can be obtained by the proposed algorithm than by basic multi-objective ant colony algorithm, and the Pareto frontier distribution is more uniform, meanwhile, the convergence and rapidity are improved.
出处
《电网技术》
EI
CSCD
北大核心
2009年第18期57-62,共6页
Power System Technology
基金
上海市重点攻关计划项目(071605123)
上海市教委科研创新项目(08ZZ92)
上海市教委重点学科建设资助项目(J51301)
关键词
多目标蚁群算法
聚类分析
PARETO最优
电网规划
multi-objective ant colony algorithm
clustering analysis
Pareto optimal
power network planning
