防灾减灾学报

针对复杂建筑环境中的火灾人员疏散路径优化问题，本研究以疏散方案的安全性与效率为核心，运用图论几何原理构建了室内火灾疏散网络。首先，结合火灾动态网络数据模型与修正后的临界簇模型（Critical Cluster Model,CCM），识别疏散网络中各节点的最大风险簇，并引入火灾风险影响因子评估风险程度；其次，构建以最短疏散时间和最小风险为目标的路径优化模型，通过改进蚁群算法的信息素管理方式和启发函数，实现时间与风险的双目标动态路径规划；最后，通过仿真实验对模型的优化性能和疏散效率进行验证。结果表明：双目标路径优化模型可有效规避高风险路径，降低危险源附近出口的使用频率，整合可用疏散资源，引导人群选择更安全、更高效的疏散路径，更适合处理建筑空间复杂的群体性火灾应急疏散问题。

This study addresses the optimization of evacuation routes for individuals during fires in complex building environments, focusing on the safety and efficiency of evacuation plans. Utilizing geometric principles of graph theory, an indoor fire evacuation network is constructed. By integrating a dynamic fire network data model and employing a modified CCM model, the study identifies the maximum risk clusters at each node within the evacuation network. A fire risk impact factor is introduced to assess the risk level of incidents. A path optimization model is developed, aimed at minimizing both time and risk. Dynamic path planning with dual objectives is achieved through improved pheromone management and heuristic functions in the ant colony algorithm. The optimization performance and evacuation efficiency of the model are validated,with results indicating that the dual-objective path optimization model can effectively avoid highrisk paths,reduce the usage rate of exits near hazardous sources,efficiently integrate available evacuation resources, and guide evacuating individuals towards safer and more efficient routes. This model is better suited to addressing the complex group fire emergency evacuation challenges in building spaces.