新安江模型参数率定的多核并行遗传算法

Multi-Core Parallel Genetic Algorithm for Parameter Calibration of Xin’anjiang Model

新安江模型参数众多,率定计算工作量大、传统率定参数精度难以满足实际生产需要的特点。本文提出了一种参数率定的多核并行遗传算法,以洪峰流量、峰现时间和洪水总量为评价目标,建立多准则参数率定模型,并采用模糊集思想把多准则问题转化为单一目标优选问题。其次针对遗传算法个体适应度计算时间较长,采用主从式并行策略实现遗传算法的并行化,提高系统计算效率。最后以我国南方地区酉水流域凤滩水库为例,该方法不仅可获得高质量的模型参数,保证模型的预报准确性,同时可解决新安江模型参数率定耗时长等问题,有效提高模型参数率定效率,为模型参数校核提供参考。

Aiming at the characteristics of many parameters of Xin’anjiang model, heavy workload of calibration calculation and difficulty of traditional calibration parameter accuracy to meet the needs of actual pro-duction, a multi-core parallel genetic algorithm for parameter calibration is proposed. Taking the peak discharge, peak time and total flood as the evaluation objectives, a multi criteria parameter calibration model is established, and the multi criteria problem is transformed into a single objective optimization problem by using the idea of fuzzy set. Secondly, in view of the long computing time of individual fitness of genetic algorithm, the master-slave parallel strategy is adopted to realize the parallelization of genetic algorithm and improve the computing efficiency of the system. Finally, taking Fengtan Reservoir in Youshui Basin in southern China as an example, this method can not only obtain high-quality model parameters and ensure the prediction accuracy of the model, but also solve the problems of long time-consuming in Xin’anjiang model parameter calibration, effectively improve the efficiency of model parameter calibration and provide reference for model parameter verification.