基于改进遗传算法的河流水质模型多参数识别

结合多种群遗传算法和自适应遗传算法,提出基于多种群-自适应遗传算法(MPAGA)水质模型多参数识别方法,实现对河流水质模型参数断面平均流速u、河流离散系数D和污染物降解系数K识别和估计。利用河流示踪剂实验监测数据,分别对美国特拉基河3个不同流量河段作水质模型多参数识别和估计。结果表明,与传统遗传算法(TGA)相比,MPAGA算法对3种不同流量河段水质模型参数估计结果准确和可靠。基于MPAGA水质模型多参数识别方法和分河段水质模型参数估计结果可提高河流水污染预测精度,判定河流水污染危险区域,为保障工农业用水安全提供重要科学依据和技术支撑。

Multi-parameter identification of river water quality model based on an improved genetic algorithm

In this study, a multiple population-adaptive genetic algorithm(MPAGA), with the integration of multi-population and adaptive genetic algorithm, is proposed to identify and estimate water quality parameters, such as water flow rate u, dispersion coefficient D and rate constant of pollutant degradation K. Then, the developed method was applied to Truckee River, America for multi-parameter identification, according to the river trace experimental data under three scenarios of different river discharges. Comparing with traditional genetic algorithm(TGA), the results demonstrate that MPAGA could obtain more accurate and reliable parameter identification results under all the three scenarios of different river discharges. The multi-parameter identification method MPAGA for water quality model and the multi-parameter identification results for different reaches, could effectively enhance pollution prediction accuracy, identify river hazardous area, and provide scientific basis and technique support for industrial and agricultural water safety.