摘要
禁忌搜索(TS)算法具有强大的全局优化性能,但其局部搜索性能易受分散性的影响;蚁群最优(ACO)算法的正反馈机制使其具有强大的局部搜索性能,但其全局优化性能的优劣在很大程度上与蒸发系数的选择有关,如选择得不合适易使算法陷于局部最优。文章将TS算法与ACO算法组合起来,提出了TS-ACO混合算法,用于求解配电网规划问题,在同时考虑扩展配电网所需的固定费用和与电能损失相关的变化费用的基础上,设计了非线性混合整数配电网规划数学模型,在一具有6个变电所、102条馈线段的配电网上进行的测试结果表明了TS-ACO混合算法的有效性。
Tabu Search (TS) behaves well in finding global optimum of combined optimization problems, whereas its local search is not satisfactory due to diversity; Ant Colony Optimization (ACO) behaves well in finding local optimum, whereas its global search depends on selection of the evaporation coefficient. An unsuitable evaporation coefficient may results in local optimum of final solutions. To compensate some limitations of single algorithms, the authors integrate the two algorithms together and put forward TS-ACO hybrid algorithm to solve the distribution network planning. On the basis of simultaneously taking account of the fixed expenditure, which is required in the expansion of distribution network and the variable cost, which is corresponding to the electrical energy losses, a non-linear hybrid integer mathematical model for distribution network planning is designed. The testing results of a distribution network which possesses 6 substations and 102 feeders show that the presented TS-ACO hybrid algorithm is effective.
出处
《电网技术》
EI
CSCD
北大核心
2005年第2期23-27,共5页
Power System Technology
关键词
配电网规划
变电所
馈线
电能损失
禁忌搜索
TS
混合算法
最优
蚁群
全局优化
Electric losses
Electric power systems
Electric substations
Mathematical models
Optimization
Planning
Problem solving