随着我国西部的大开发和“一带一路”的建设，迫切需要大力发展交通、矿山、水利等重大“生命线”工程，而深埋长大隧道往往是这些生命线工程的关键控制性工程。限于地质、地形和自然环境条件，从施工工期、造价和技术进步3个方面考虑，TBM(Tunnel Boring Machine)工法是深长隧道开挖的优先选择。深长隧道TBM的适应性受到众多因素的影响，难以进行有效和定量的评价。主要影响因素为不良地质，如突涌水、软岩大变形、断层破碎带、岩爆等;此外，隧道的设计、隧址地质条件等对TBM的选型也有重要影响。人工智能方法具有能够分析复杂因素影响和处理复杂问题的突出特点，可用于TBM选型适应性的有效评价。首先，基于层次分析法和模糊综合评判方法，通过TBM选型评价知识的获取，选取能够充分反映不同机型地质适应性差异、具有代表性和区分度高的7个评价指标，构建了TBM选型适应性评价指标体系及模糊综合评价模型，确定了各个评价指标的模糊隶属函数。其次，通过编写权重辅助计算程序，确定了3种TBM机型选型适应性评价指标的权重;其中，为了避免单指标决策的局限性和主观臆断的缺陷，采用智能设计理论和决策理论相结合的方法，完成了多指标智能决策的定量化选型。将评价模型与知识获取相结合，以规则的形式表示知识，构建了TBM选型适应性评价知识库。最后，基于智能评价决策支持系统平台IDSDP，开发了TBM选型适应性智能评价决策支持系统，为深长隧道TBM选型提供了一种新的量化评价方法。利用该系统对高黎贡山铁路隧道TBM的选型进行了适应性评价，评价结果与实际情况相吻合。

With the construction of the Western region of China and the development of “ the Belt and Road Initia- tives”,there is an urgent need to develop some major “lifeline” projects such as transportation,mining and water con- servancy. Deep buried tunnels are often the key control projects for these lifeline projects. Limited to geological,topo- graphical and natural environmental conditions,the TBM method is the preferred choice for deep tunnel excavation in terms of construction period,cost and technical progress. The adaptability of deep tunnel TBM is affected by many fac- tors,making it difficult to conduct an effective and quantitative evaluation. The main influencing factors are unfavorable geology,such as inrush water,large soft rock deformation,fault fracture zone,rock burst,etc. In addition,the design of the tunnel and the geological conditions of the tunnel site also have an important impact on the selection of TBM. The artificial intelligence method has the outstanding features of being able to analyze the influence of complex factors and dealing with complex problems,and can be used for the effective evaluation of the adaptability of TBM selection. First- ly,based on the analytic hierarchy process ( AHP) and fuzzy comprehensive evaluation method,through the acquisi- tion of TBM selection knowledge,seven indicators which can fully reflect the differences of adaptability,representative- ness and high discrimination are selected. The evaluation index system and fuzzy comprehensive evaluation model of TBM selection adaptability are constructed,and the fuzzy membership function of each evaluation index is determined. Secondly,the weights of three TBM model selection adaptability evaluation indicators are determined by compiling the weights assistant calculation program. Among them,in order to avoid the limitation of single index decision and the de- fect of subjective assumption,a combination of intelligent design theory and decision-making theory is adopted to carry out the quantitative selection of multi-indicator intelligent decision-making. Combining the evaluation model with knowledge,expressing the knowledge in the form of rules,the TBM selection adaptive evaluation knowledge base is constructed. Finally,based on the IDSDP intelligent evaluation and decision system platform,an intelligent evaluation and decision-making system for TBM selection is developed,which provides a new quantitative evaluation method for the deep tunnel TBM selection. The adaptability of TBM selection for Gaoligongshan Railway Tunnel is evaluated by u- sing the decision-making system,and the evaluation results are in good agreement with the actual situation.