多无人机协同路径规划计算资源分配方法研究

Research on Resource Allocation method for Cooperative Path Planning of UAV

摘要:

针对多目标多机协同路径规划问题，在MOEA/D算法求解的基础上，采用深度强化学习方法对MOEA/D算法中计算资源分配方法进行了研究；对多机协同路径规划问题进行了研究，分析了相关约束以及优化目标，建立多机协同路径规划多目标优化模型；结合协同进化思想，对基于分解的多目标进化协同路径规划进行了研究；对基于强化学习的计算资源分配策略进行了研究，实现了深度强化学习在多目标优化计算资源分配问题上的应用；实现了多机协同路径规划仿真验证；经仿真测试，算法以更高性能完成多机协同路径规划任务，提高了多机协同路径规划中计算资源分配策略的能力。

Abstract:

Aiming at the multi-objective multi-UAVs cooperative path planning problem, based on the solution of MOEA/D algorithm, a deep reinforcement learning method is adopted to study the computational resource allocation method in MOEA/D algorithm; The multi-UAVs cooperative path planning problem is studied, the relevant constraints and optimization objectives are analyzed, and the multi-objective optimization model of multi-UAVs cooperative path planning is established; Combined with the idea of co-evolution, the multi-objective evolutionary cooperative path planning based on decomposition is investigated; the computational resource allocation strategy based on reinforcement learning is investigated, and the application of deep reinforcement learning in the multi-objective optimization of computational resource allocation is realized; The simulation verification of the multi-UAVs cooperative path planning is realized; The algorithm completes the multi-UAVs cooperative path planning task with a higher performance after the simulation test and the performance of the computational resource allocation strategy in the real problem is improved; The performance of the computational resource allocation strategy in the problem is improved.