基于混合多目标进化算法的多无人机侦察路径规划

由于侦察任务的复杂性和不确定性，无人机对其目标的侦察时间往往是不确定的。将多无人机对观测时间不确定目标的侦察路径规划问题建模为使任务时间、编队总耗时和编队规模同时最小化的多目标优化路径规划问题。对此，在基于ε 占优的稳态多目标进化算法基础上引入多目标局部搜索，给出了混合ε 占优多目标进化算法,提出了一种使用插入最近点方法的启发式遗传操作。实验结果表明，算法能够有效解决所研究的问题，并且其优势随着问题规模的增大而显著。

Multiple UAVs routing in reconnaissance mission based on hybrid multi-objective evolutionary algorithm

The observation time on the target is usually uncertain due to the complexity and uncertainty of reconnaissance missions. The multiple unmanned aerial vehicles (UAVs) reconnaissance problem with a stochastic observation time (MURSOT) is modeled as a multi-objective optimal routing problem including minimizing mission duration, total time and fleet size. For solving this problem, a multi-objective local search is incorporated to a steady-state multi-objective evolutionary algorithm (MOEA) with ε-dominance conception (epsMOEA). Besides, several heuristic genetic operations using the insert-to-nearest method (INM) are proposed. Experimental results show that the proposed method is effective on MURSOT and its superiority is more remarkable with the growth of the size of missions.