基于改进DQN算法的海上搜救路径规划方法OA
Path Planning Method for Maritime Search and Rescue Based on Improved DQN Algorithm
针对海上复杂环境下多任务点的搜救策略和路径规划问题,论文提出了一种基于改进DQN算法的海上搜救路径规划方法.该算法构建融合路径长度、任务完成时效及环境安全等指标的多目标奖励函数,同时在经典DQN框架基础上能够提升路径决策稳定性,平衡训练中探索与利用.实验结果表明,该算法在训练过程中能够规划出更高效的海上多任务点搜救路径,表现出良好的收敛性、稳定性和效率,为保障海上搜救任务稳定性、高效执行提供支持.
To address the challenges associated with search and rescue strategies and path planning for multiple task points in the complex marine environment,this paper proposes a maritime search and rescue path planning method based on improved DQN algorithm.The algorithm constructs a multi-objective reward function that integrates indicators such as path length,task completion timeliness,and environmental safety.Additionally,it can improve the stability of path decision-making and balance exploration and utilization during training based on the classical DQN framework.The experimental results demonstrate that the proposed algo-rithm can plan more efficient maritime multi-task point search and rescue paths during training,exhibiting good convergence,sta-bility,and efficiency,thereby supporting the stable and effective execution of maritime search and rescue missions.
闫宇轩;李瑞;陈行军
海军大连舰艇学院 大连 116018海军大连舰艇学院 大连 116018海军大连舰艇学院 大连 116018
信息技术与安全科学
强化学习DQN算法海上搜救路径规划
reinforcement learningDQN algorithmmaritime search and rescuepath planning
《舰船电子工程》 2026 (3)
12-15,50,5
评论