基于实码遗传算法的河流水质模型的参数估计

针对理想条件下采用解析法往往导致水质参数估计较大误差的问题，以及针对含有多参数的二维水质模型的参数估计问题，具体介绍了采用有限单元法和实码遗传算法求解二维水质模型未知参数的基本步骤，对水质参数(包括纵向、横向弥散系数和衰减系数)分别进行编码，通过计算机模拟浓度输出，并与实测值比较从而得出最优的水质参数估计值．算例表明，采用有限单元法-遗传算法估计河流水质模型的参数是可行的．

Parameter Estimation of River Water Quality Model Based on Real-Coded Genetic Algorithm

Because of the considerable errors caused by applying the analytic method for the parameter estimaton of water quality under ideal conditions, and with a view to solving the problem of several parameter estimations in two-dimensional water quality model, this paper presented the basic steps for the parameter estimation of the two-dimensional water quality model, based on finite element method and real-coded genetic algorithm. The method involved coding the parameters, including longitudinal dispersion coefficient, transversal dispersion coefficient and decay coefficient, and producing the concentration outputs. The results were compared with the actual monitoring data, and the optimal parameter estimating values were obtained. The example showed that this approach of parameter estimation, based on the finite element method and genetic algorithm, is applicable.