动力测桩中完整桩多参数反演问题的遗传算法  

A Genetic Algorithm for Multi-parameters Reverse Deduce of Integrated Pile in Dynamic Testing

作　　者：陈建功[1] 张永兴[1] 郑永保[1] 

机构地区：[1]重庆大学土木工程学院,重庆400045

出　　处：《重庆大学学报（自然科学版）》2003年第11期126-129,共4页Journal of Chongqing University

摘　　要：动力测桩中需要同时对桩长、波速等多个参数进行反演估算,这是个较困难的问题。为此,运用遗传算法的基本原理,提出一种动力测桩中完整桩多参数反演问题的遗传算法。动力测桩多参数反演问题可归结为一个非线性优化问题,采用传统的非线性优化方法很容易使目标函数处于局部极小域,很难取得较好的效果,而利用遗传算法具有自组织、自适应和自学习等智能特性,可以得到满意解,且收敛速度较快,利用Matlab工具箱进行了实例计算,表明此法既简单精度又高,是一种行之有效的计算方法。In pile dynamic testing, the parameters such as pile length, wave velocity must be calculated, and it is a difficult problem. Multiparameters reverse deduces for integrated pile in dynamic testing is a nonlinear optimization problem. It is not easily solved by traditional optimization methods because the problem is a convex programming. These methods often make the target function be placed in secondly optimization zones and can not get the good result. On the contrary, the genetic algorithm is a new optimization method which mimic the nature evaluation and can solve the problem accurately and quickly. Based on the main principle of the genetic algorithm and the properties of selforganizing, selfadaptation and selfstudy, the authors put forward a genetic algorithm to solve the problem of multiparameters reverse deduce for integrated pile in dynamic testing. With the toolbox of Matlab, several examples are calculated. The method is simple and precise.

关 键 词：桩基 遗传算法 反演 动力测试 

分 类 号：TU432.3[建筑科学—岩土工程]