基于局部路径重连的多星测控调度遗传交叉算子

Genetic Crossover Operator Based on Local Path Relinking for Multi-Satellite TT&C Scheduling

为提高多星测控调度问题简单遗传算法的搜索精度,设计一种基于局部分层路径搜索的交叉算子(local layering path-relinking crossover operator,LLPRCO)。分析多星测控调度问题的遗传算法编码特点,得出解空间的复杂性和基因间较弱的相关性。借鉴路径重连的思想,利用构成初始解和引导解要素的差异性,构建从初始解出发的分层搜索邻域,将邻域中满足模型约束的解作为交叉的结果,并创建2个仿真场景进行验证。仿真结果表明,该交叉算子与具备一定随机特性的交叉算子混合使用能明显提高遗传算法对问题的求解性能。

For improving the genetic algorithm searching precision used in the multi-Satellite TT＆C （tracking telemetry and command） scheduling problem, a directed crossover operator based on local path layering relinking is proposed. The solution space complexity and weak correlation between genes are found by analyzing the GA encoding character. Based on path relinking, use the difference between initial solution and guiding solution to construct the layering neighborhood of initial solution, and take the solution in the neighborhood that meets the constraints of model as the crossing result, which is based on the idea of path relinking, then constructing two simulation cases to verify the method. Simulation result indicates that the proposed operator combined with a crossover operator possessing randomness and can improve algorithm performance significantly.