一种基于分散搜索的多星测控调度遗传算法

多星测控调度是一个具有大搜索空间的多峰问题。针对简单遗传算法求解易陷入局部最优和不稳定的缺陷,借鉴分散搜索多样化采样、局部寻优的特点,提出一种基于分散搜索的混合遗传算法,在全局的随机搜索中嵌入全局的定向搜索。在描述问题的基础上,提出可进行细粒度搜索的可行解表示方式,构建算法的整体流程,并设计由输入参数控制的多样化初始集产生方法、基于质量和多样性原则的参考集生成和更新方法、吸取被组合个体优良成份的解组合方法及基于启发式局部搜索的解提高方法等算法要素。仿真表明新算法在求解质量上比简单遗传算法有明显提高。

A Genetic Algorithm Based on Scatter Search to Multi-satellite TT&C Scheduling

Multi-Satellite TT&C scheduling is a multi-peak problem with huge search space.The simple genetic algorithm solving is prone to get into local optimization and instability.Because Scatter Search can sample diversifiedly and optimize locally, a hybridized genetic algorithm based on scatter search was proposed, which embeds the global directed search in the global stochastic search. After the problem was described,the representation of feasible solution was designed,which was convenient for searching roundly, then the process of the algorithm was construced, and the main elements of the algorithm were presented, which includes diversification generator controled by input parameters , reference set generating and updating method based on quality and diversification principle, the combination method drawning on the good components of combinated solutions, and the improvement method based on heuristic local searching. Simulation result shows the new algorithm can improve the quality of the solutions,compared with the simple genetic algorithm.