紧急处置与救援是煤矿安全生产的重要保障之一，救援路径优选的有效预判，能降低其可能造成的人员、生产、经济、社会等多重损失。本文通过对事故相关因素作预警分析，综合群智能优化算法、多目标优化技术、模糊层次分析法，研究煤矿灾变条件下井下应急救援路径优化，主要内容包括： 研究了基于模糊AHP的井下多目标蚁群应急救援路径优化方法。通过对事故相关因素作预警分析，构建基于模糊AHP的多目标蚁群路径优化方法。通过现场调研、专家经验等方式遴选主要因素，建立事故树-模糊AHP层次结构模型。利用标度比较法和“专家系数法”形成基于信心指数法的主要指标成对比较专家判断矩阵；综合利用判断矩阵的几何平均特征、算术平均特征以及特征向量，采用“三级滤波法”逐层求解，确定充分反映判断矩阵特性的指标权值，结合蚁群算法求解应急救援最优路径。为了对方法的应用效果进行验证，将其用于井下8个节点、12条边的网络拓扑图救援路径综合预判模拟。实验结果说明，模型能够计算出各个影响因素的权值，验证了本文方法的有效性，同时和巷道当量长度法的最优路径及解的收敛性相比较，该方法最优解的收敛性优于巷道当量长度法。 研究了基于模糊AHP模型的群智能井下救援路径综合评判方法。针对路径优化中蚁群算法的参数依赖问题，采用粒子群算法的全局搜索功能，优化α、β、ρ三个重要参数，构建出基于粒子群的AHP多目标信息熵蚁群路径优化算法。利用巷道行走难度、瓦斯浓度、一氧化碳浓度、风量风速和灾变情况五个主要因子的量化值，作为节点信息熵，更新蚁群算法信息素。根据多因素信息素更新寻找最优路径，并对过程的收敛性进行判断。实验应用于煤矿矿井内的8个节点、12条边的网络拓扑结构，用该方法确定实例的最优救援路径，实验结果表明方法对于救援路径优化选择的有效性，同时解的适应度值表明最优解具有较好的收敛性。 研究了基于多目标融合的蚁群井下救援路径优化方法。通过对各相关影响因子的分析，提出多目标融合“地势图”的蚁群路径优化方法。首先利用基于滑动时间窗的ICA数据降噪法，对融合数据进行预处理，由多目标因子建立多级地势图，作为蚂蚁选择路径的集合；然后蚂蚁分别通过各级地势图并释放信息素，大小由地势图节点上融合因子的值决定；最后将各级地势图对应边上的信息素进行归一化叠加，即多目标融合信息素更新规则，实现蚁群路径优化选择。实验中，将其应用于21个节点、26条边的井下网络拓扑结构，结果说明该方法在解的适应度变化趋势上具有较好的特性。 研究了井下应急救援模糊AHP模型的群智能优化方法应用方案。制定煤矿井下应急救援路径选择实验的总体方案，应用于井下31个节点、43条边的网络拓扑结构，通过得到的最优路径及其运行时间及收敛性，验证了方法用于井下应急救援的有效性和适用性。最后将方法应用到实际煤矿井下巷道299个节点、438条边中，在灾害节点动态发生的情况下，建立群智能动态井下应急救援路径优化策略。 本文针对煤矿井下应急救援的特点，基于模糊AHP多目标模型提出群智能路径优化方法，对于煤矿的应急救援安全路径评判表现出良好的寻优能力和收敛性，为井下应急救援的路径选择提供了新的智能化方法支持。

Emergency rescue is one of the important guarantees for coal mine safety production, and the effective prediction of the optimal selection of the rescue route can reduce the multiple losses such as personnel, production, economy and society. Through the analysis of accident related factors and comprehensive utilization of swarm intelligence optimization algorithm, multi-objective optimization technology and fuzzy AHP method, this paper studies emergency rescue route optimization in coal mines. The main contents include: The method of route optimizing for multi-objective ant colony emergency rescue based on fuzzy AHP is studied. Through the analysis of accident related factors, a multi-objective ant colony optimization method based on fuzzy AHP is constructed. Through the field investigation and expert experience, the main factors are selected, and the fault tree fuzzy AHP hierarchical structure model is formed. By using the scale comparison method and the expert coefficient method, the expert judgment matrix of pairwise comparison is formed based on the confidence index method. Comprehensively utilizing geometric mean, arithmetic mean feature and feature vector of judgment matrix, the “three level filtering method” is adopted to solve the problem layer by layer to determine the weights of the indexes that fully reflect the characteristics of the judgment matrix. At the same time, ant colony algorithm is used to solve the optimal path of emergency rescue. In order to validate the application effect of the method, it is used in the comprehensive prediction simulation of rescue path for 8 nodes and 12 edges of coal mine network topology. The experimental results show that the weight of the model is consistent with the expert experience, and the validity of the method is verified. At the same time, the convergence of the optimal solution is better when compared with those of the roadway equivalent length method. The comprehensive evaluation method of underground mine rescue route based on fuzzy AHP model is studied. In view of the parameter dependency problem of ant colony algorithm in path optimization, three important parameters of α, β, ρ are optimized by using global search function of particle swarm algorithm. Based on particle swarm optimization, an ant colony optimization algorithm of multi objective AHP information entropy is constructed. Using the quantization values of the difficulty of walking, gas concentration, carbon monoxide concentration, air volume & wind speed and catastrophic situation as node information entropy, the pheromone of ant colony algorithm is updated. According to the multifactor pheromone update, the optimal path is searched and preserved, and the convergence of the process is judged. The experiment is applied to 8 nodes and 12 edges of coal mine network topology, and the optimal rescue path is confirmed by this method. The experimental results show the effectiveness of the proposed method in the optimization of the rescue route, and the fitness of the solution shows that the optimal solution has a good convergence. The ant colony route optimization method of underground rescue based on multi-objective fusion is studied. Through the analysis of relevant factors, an ant colony optimization method based on multi objective terrain map is proposed. Firstly, the ICA data noise reduction method based on sliding time window is used to preprocess the fused data. A multi-level terrain map is constructed from multiple objective factors as a set of ant choice paths. The ants then release the pheromone at each level, and the size is determined by the value of the fusion factor on the topographic map. At last, the pheromone on each side of the terrain map at different levels is normalized and superimposed. That is, the multi-target fusion pheromone updating rules are adopted to realize the ant colony routing optimization. In the experiment, it is applied to 21 nodes and 26 edges of coal mine network topology, and the result shows that the method has better characteristics in the adaptability of the optimal solution. The application scheme of swarm intelligence optimization method for fuzzy AHP model of underground emergency rescue is studied. The overall scheme of coal mine emergency rescue route choice experiment is worked out, which is applied to 31 nodes and 43 edges of coal mine network topology. Through the optimal path length, running time and convergence, the validity and applicability of the method used in underground emergency rescue are verified. Finally, the method is applied to 299 nodes and 438 edges of coal mine network topology, and it is proved that the method has the ability to produce optimization strategy of dynamic emergency rescue route in the case of dynamic occurrence of disaster nodes. Aiming at emergency rescue in coal mine, this paper proposes swarm intelligence path optimization method based on fuzzy AHP and multi objective model. The safety path evaluation shows good optimization ability and convergence, and a new intelligent method is put forward for the path selection of underground emergency rescue.