In post-disaster rescue and relief assurance,emergency water supply plays a crucial role.However,due to the complex and changeable mountainous terrain and geomorphic conditions,on-site command and dispatch are particularly critical in determining whether rescue workers can quickly de-ploy support equipment for emergency water supply operations.Based on the multi-agent proximal policy optimization(MAPPO)algorithm,a path planning system was designed and experimental verifi-cation was conducted.The feasibility of the path planning system is confirmed according to the results of the reward diagram,and the system operation was visualized,demonstrating that the path planning system could preliminarily meet the requirements for path planning of emergency water supply equipment in mountainous areas.On this basis,Mask2Former's image segmentation model was inte-grated to optimize the path planning system for mountainous area emergency water supply equipment.By combining the results of ground object information output with path planning results,significant fluctuations were avoided in the results of single path planning algorithms when affected by the environ-ment,thereby enhancing the robustness and reliability of path planning.Integrating this path planning system into the command platform for mountainous area emergency water supply equipment solved the path planning issues in mountainous,providing strong support for the actual operation of emergency water supply equipment in mountainous regions.
关键词
路径规划/应急供水/强化学习/指挥调度/多智能体强化学习算法
Key words
path planning/emergency water supply/reinforcement learning/command dispatching/multi-agent proximal policy optimization algorithm
维普
万方数据