基于遗传算法的多无人机协同侦察任务规划研究

Research on Multi-UAV Cooperative Reconnaissance Mission Planning Problem Based on Improved Genetic Algorithm

通过对侦察任务特性的分析，充分考虑了多UAV协同侦察任务规划问题中的多个约束条件和性能规划指标，特别是对目标侦察次数、侦察时间和对UAV类型的要求以及对UAV性能的约束，基于建模理论，建立了更加贴近军事应用实际的多基地多目标多UAV协同侦察问题的数学模型。并针对模型存在的NP难和复杂约束等问题，基于经典遗传算法的特点，设计了适用于该问题求解的改进的遗传算法，利用启发式插入算法构造出的初始可行解在很大程度上避免了进化过程收敛太慢的问题。

Through the analysis of characteristics about reconnaissance missions, this paper takes plenty of constraints and performance planning targets in multi-UAV cooperative recon- naissance mission planning problems into full consideration, in particular the requirements of targets surveillance times, reconnaissance time, types of UAV as well as the performance re- strictions of UAV. And thus, based on modeling theory, this paper establishes a more practical mathematical model for the multi-base, multi-target, muhi-UAV cooperative reconnaissance problem. Moreover, aimed at the problems existed when dealing with the model, such as NP- hard and complex constraints, this paper has designed a modified genetic algorithm, based on the classical one, and meanwhile constructed the initial feasible solutions through a heuristic method, which would avoid the over-slow convergence in the process of evolutionary to a large degree.