克隆选择算法在基坑支护位移反分析中的应用

在基坑工程中,传统的位移反分析方法计算量较大,不易掌握,且求得的岩土体力学性能参数的精度较低。为此,提出应用克隆选择算法求解位移反分析问题的新方法,其核心是采用BP网络为正演工具代替有限元方法,描述土体力学性能参数和位移的非线性映射,可减少计算量,并利用克隆选择算法为反演工具,通过反演分析确定土体力学性能参数,具有较高的精度。依据该反分析方法获得的岩土体力学性能参数又可以估算下一工况的基坑位移。仿真试验结果表明了新方法的可行性和有效性。

Application of clonal selection algorithm for foundation pit displacement back-analysis

Based on the fact that conventional displacement back-analysis methods had disadvantages such as complicated compution,difficult to catch,and low precision solutions,a new method was proposed to solve displacement back-analysis problem.Its emphases were that BP nets replaced plastic FEM program as a tool in the forward process to map the non-linear relations of samples and to reduce computing.Clonal selection algorithm was used as tool in the backward process to optimize index parameters.At the same time,high prcision displacement can be achieved based on high prcision index parameters.Simulating results show that the proposed method is not only feasible but also effective.