基于Multi-Agent的矿井人员疏散数值模拟

煤矿井下发生灾害时,井下应急疏散建模仿真可以验证疏散方案是否合理,从而为救灾逃生提供合理指导。目前在煤矿井下人员疏散研究中,对因个体差异导致不同疏散结果的认识和建模尚有欠缺,并且对不同路权评估方法下最低花费路径的选取也鲜有提及。针对煤矿井下出现安全事故时人员的安全疏散问题,基于Multi-Agent与Dijkstra算法,利用RepastJ3.0开发平台,建立了井下人员安全疏散模型,探讨了井下工作人员对巷道网络拓扑关系的认知差异性对疏散效果的影响,以及不同路权评估方法下最低花费路径的选取策略。疏散模型仿真结果表明,该模型能够直观呈现不同类型井下工作人员的动态疏散过程,可以为井下应急疏散方案的制定提供理论支持。

Simulation of Evacuation in Underground Coal Mines based on Multi-agents

Computer simulation of mine evacuation can provide guidance for making emergency evacuation plans and verify the rationality of the plans in case of mine accidents or disasters. At present, the present studies seldom take account of individual cognitive differences and path selection differences that should be given weighted treatments. In this paper, a coal mine evacuation model is developed based on Multi-Agent and Dijkstra algorithm using RepastJ3.0 Java development platform, and the effects of the roadway topology knowledge level of the miner agents and the roadway variety on safety evacuation is discussed with the proposed model. The results indicate that safety evacuation varies with the roadway topology knowledge level of the miner agents and road weighted settings. The proposed model can directly and dynamically present the simulation process, which could be useful to provide theoretical supports for making emergency evacuation plans.