情绪驱动下的火灾疏散路径规划

——基于MADDPG算法的分析

打开文本图片集

Abstract：Existing evacuation models sufer from issues such as high complexityand inaccurate quantification of psychological andbehavioral factors，failing tomeetreal-time requirements.Toaddress this，anemotion-driven fire evacuation path planning method based on the MADDPG algorithm is proposed.An emotional state model for fire escape scenarios is constructed，analyzing theexit selectionbehaviorof non-panickedand panicked evacuees to establish a behavioral characteristics model of evacues.The MADDPG algorithm is employed to design areward function，assigning differentrewards or penalties for various actions to achieve inteligent agent evacuation path planning.Theresults show that the proposed method plans smoother paths.When thenumberof evacuees is 2Oo，the evacuation time is 92.3 seconds，with a higher average reward value and beter convergence performance.This method achieves dynamic path optimization byquantifying the impact of emotions on evacuation behavior，providing a new perspective forbuilding more realistic and efficient intelligent evacuation systems.

Keywords： emotion-driven；MADDPG algorithm； fire evacuation；path planning

0引言

体及工业厂房数量激增，火灾事故的复杂性和破坏性显著提升[]。（剩余7252字）