自适应遗传算法在某拱坝渗流参数反演中的应用

Application of adaptive genetic algorithm in the inversion of the seepage parameter of some arch dam

利用水头实测资料,以渗透系数为待反演参数,在采用基本遗传算法进行参数反演研究的基础上,针对简单遗传算法难以确定交叉率和变异率的最佳值及计算量较大、易早熟等缺点,提出以自适应遗传算法来解决工程中的这类反演问题;为力求使改进的遗传算法计算量更小,收敛性更强,同时结合二维稳定渗流有限元工程实例进行了反演计算.结果表明,自适应遗传算法在保持简单遗传算法优点的同时,有效地提高了算法的收敛性,并在一定程度上克服了简单遗传算法的早熟问题.

The permeability coefficient of the arch dam is inversed based on the actually measured water-head data.Because basic genetic algorithm has the disadvantages that the determination of the optimal values of crossover rate and mutation rate is difficult,computation quantity is great and premature convergence easily appear,it is proposed that adaptive genetic algorithm is used for this propose.The adaptive genetic algorithm is used for the inversion of the permeability coefficient of a 2D steady seepage finite element engineering case.The result shows that the adaptive genetic algorithm not only remains the advantages of the basic genetic algorithm but also overcomes its disadvantages.