基于蚁群算法的地铁隧道沉降预测研究

紧邻大型深基坑的地铁隧道因影响因素极其复杂，难以准确预测其沉降变形，以致无法较早判定隧道的安全状况，亦无法根据其变形信息及时、有效地指导基坑施工。针对这一问题，立足TSP基本蚁群算法模型，结合地铁隧道沉降变形实际，在确定模型路径选择机制的基础上，通过合理构建信息函数和启发函数，确立信息素更新机制以及蚂蚁搜索机制，最终建立了地铁隧道沉降预测的蚁群算法模型，并用实例验证了模型的预测效果，为运营中的地铁隧道特别是紧邻大型基坑的地铁隧道沉降预测开辟了一条新的途径。

Study on Settlement Prediction of Metro Tunnels Based on Ant Colony Algorithm

Due to the extremely complex factors, it is very difficult to predict the settlement and deformation of Metro tunnels close to large scale deep foundation pits. Therefore, it is impossible to determine the safety conditions of the Metro tunnels and it is also impossible to provide guidance for the construction of the foundation pits. Ant colony algorithm model is established on basis of TSP basic ant algorithm model, actual settlement and deformation of Metro tunnels, determination of the path of the ant colony algorithm model, establishment of information function and heuristic function, and determination of pheromone updating mechanism and ant searching mechanism, so as to predict the settlement of Metro tunnels. The effect of the ant colony algorithm model is validated in engineering cases. The ant colony algorithm model provides a new way for the prediction of the settlement of operating Metro tunnels, especially those close to large scale deep foundation pits.