基于GA的感潮河流水污染控制多目标优化

Multi-objective Programming for Water Pollution Control in a Tidal River Based on GA

文章针对感潮河流水质时空变化的特点,建立了污水处理费用函数、动态水质模型与NSGA-Ⅱ算法相耦合的流域水污染控制多目标优化方法。采用该方法对深圳河流域5个污水处理厂的污染物去除效率进行优化,得到流域污水处理费用与河流上、中、下游控制断面水质潮周期超标率的多目标解集排列图。结果表明:(1)所有控制断面水质超标率都小于50%、25%和0%时,最小处理费用分别为1.92、1.94和1.98亿元/a;(2)上、中、下游控制断面各自的水质超标率为0%时,最小处理费用分别为1.98、1.91和1.90亿元/a。决策者可根据水质管理要求和经济条件在多目标解集排列图中进行方案比选。该方法的应用可为感潮河流水污染控制提供科学依据。

This paper deals with a case study about catchment of the Shenzhen River,a tidal river,within which there are five wastewater treatment plants（WWTPs）.With the objectives to curtail the total operation cost of the wastewater treatment and optimize the pollutants removal rates of the plants,a multi-objective optimization approach,integrated with cost function of WWTPs,a dynamic water quality model and Non-dominated Sorting Genetic Algorithm II（NSGA-II）,was developed.Results suggested various schemes for achieving minimum operation cost of the WWTPs with different non-compliance rates of the water quality in different control river＇s sections in the upper,middle and downstream reaches respectively.Thus,the study offers a variety of scenarios for the pollution control of the river catchment according to the existing economic and technical circumstances as well as the space and temporal distributions of the river＇s water quality.