基于改进遗传算法的河流水污染源反演方法

Since the discharge history of pollution source is hard to master for the first time, an IGA based pollution source inversion method is developed to identify the release history of an unknown pollution source occurred upstream by integration with multi-population and adaptive Genetic Algorithms using one-dimensional river water quality model and related water quality monitoring data. The developed method is then applied to three trace experiments under various river discharges in order to identify the release history of the pollution source in Truckee River, America. The results demonstrate that the developed IGA based pollution source inversion method can significantly identify the release histories of all three trace experiments. Errors of the inversion results for three trace experiments are all within acceptable limits. Meanwhile, the developed IGA based pollution source inversion method also can guarantee the reliability and stability of the pollution source inverse, obtain satisfactory release histories under all three race experiments with different river discharges, and provide scientific and technical supports for accurate traceability and management of river environmental pollution.