蚁群算法在泵站单机组优化运行中的应用

Application of ant colony algorithm to optimal operation of single pump unit at pumping stations

蚁群算法在一些组合优化调度问题中已有成熟的应用,根据这些问题与同属组合优化调度问题的泵站优化运行的相似性,结合蚁群算法的特点合理地设定目标函数和约束条件,建立了泵站单机组优化运行的蚁群算法数学模型,提出了求解该模型的蚁群搜索方法。通过构造基于问题成分的有向赋权图,用人工蚁搜索寻找问题的可行解集,利用状态转移规则和信息素更新逐步逼近最优解。经过实例计算并与同等离散情况下的动态规划法和商业软件进化求解算法的计算结果比较,蚁群算法的计算结果更优,计算用时更少,表明该算法在泵站优化运行及相近领域有较高的实用价值。

An ant colony algorithm （ACA） model for optimal operation of pumping unit is proposed considering the similarity of optimal operation of pumping stations to other combination optimal scheduling problems of mature ACA applications, and its solution method of ants searching with the well-established object function and constraints is presented. First, feasible solutions are constructed by iteratively searching of artificial ants after converting the problem of minimum cost seeking into the one of shortest path searching, and then an optimal solution is obtained by the rule of state transition and pheromone updating. Calculation of a practical case and comparison with the dynamic programming and evolutionary solving method in commercial software under the same discrete conditions, shows less computational cost and better results of the present ACA. Thus, it provides a valuable tool for optimal operation of pumping stations and for other related fields.