遗传算法在水文地质参数反演中的应用  

Application of Generic Algorithm to the Inversion of Hydrogeological Parameters

作　　者：魏连伟[1] 邵景力[2] 张建立[3] 崔亚莉[2] 

机构地区：[1]天津大学,天津300072 [2]中国地质大学,北京100083 [3]中国水利水电科学研究院遥感中心,北京100044

出　　处：《工程勘察》2004年第3期28-31,34,共5页Geotechnical Investigation & Surveying

基　　金：国家自然科学基金资助项目 (40 2 72 10 5 )

摘　　要：本文基于遗传算法 (GA)这一全局优化技术 ,以及地下水模拟的有限元模型 ,给出了地下水系统的反演方法。并应用到北京市应急水源地水文地质参数反演中 ,反演参数包括渗透系数和给水度 ,共 8个分区的 2 4个变量 ,来验证模拟模型的可用性、适用性和鲁棒性。结果表明 ,一般情况下 ,遗传算法可以得到比较满意的结果 ,考虑测量误差和减少观测孔个数的情况下也可以得到与"标准"值相近似的解。通过优化各个变量区间的加速遗传算法可以提高算法的收敛速度。Based on the global optimization technique of generic algorithm(GA) which is coupled with the finite element model for the simulation of groundwater,the inverse modeling for groundwater systems is presented.GA is applied to inverse the hydrogeological parameters for eight zones of unconfined aquifer in Beijing.The parameters include the coefficient of permeability and the coefficient of storage(24 variables in eight zones) to verify the applicability,adequacy and robustness of the simulation modeling.The results show that GA is satisfactory to the calculation.Under the conditions of the measuring error and the reduction of observation wells,the solution that is similar to the standard value can be obtained.Through the optimization for the ranges of variables the accelerating GA can improve the convergence of GA.When the obtainment of field parameters is different GA can be used in the inversion simulation of aquifer.

关 键 词：遗传算法 水文地质参数反演 有限元 地下水模拟 

分 类 号：P641.2[天文地球—地质矿产勘探]