多参数遗传算法在地铁穿越工程中的应用

复杂地铁穿越工程中既有地下结构的沉降和变形预测是一个非常困难和复杂的问题。针对传统正分析方法在穿越工程中预测精度低的弊端，提出了一种全新的基于多参数自适应遗传算法的反分析预测模型。该模型是一种考虑施工过程的动态反分析方法，它能够综合考虑之前两个施工步序的影响，反演得到最佳的地层参数。运用该预测模型在北京地铁四号线下穿施工中对既有二号线宣武门站的结构沉降进行预测，结果表明：①引入自适应机制的该遗传算法能够改善多参数预测时局部收敛的现象，使得反演参数的“最优性”大大增加；②与其它预测方法相比，该方法的适应性更强，预测精度更高，值得应用和推广。

The application of multi-parameters genetic algorithm in the subway crossing project

 The settlement and deformation prediction of the existing station has always been a very difficult and complicated problem in the subway crossing project. Facing the low accuracy of the positive analysis method, come up with a new back analysis model, which is based on the adaptive genetic algorithm of multi parameters. This model can comprehensively consider the previous two construction steps, and get the best physicomechanical parameters of soil. Based on the new prediction model, the subsidence prediction results of the existing Xuanwumen station by the Beijing metro line 4 under-crossing the Beijing metro line 2 shows that: ①The adaptive genetic algorithm can improve the performance of local convergence, which makes the back-analysis soil parameters more and more accurate; ②Compared with the other methods, the new prediction model has a better adaptability and precision, which should be promoted and applied in projects.