改进GA法在水污染控制系统多目标规划中的应用

本文提出应用一种改进GA法求解水污染控制系统中存在的多目标规划问题，该算法采用十进制编码原理，所有操作均在实数域内进行，从而避免了二进制中编码和解码的麻烦．同时引入变权数综合距离评判模型和多个体交叉机制，对多目标函数进行求解．通过实例研究表明，改进的GA法得到最小适应度函数值在0．0468，0．0567]之间，当自变量为(0、0337，0．5663，0．4，0，0，0．4，0．1373，0，0．5，0，0．3，0，0．2154，0．1346)。时达到最优、所对应的目标函数分别是：工程费用函数312．785万元，地下水总用量0．30125m^3/s和河流断面氧亏值6．69426mg／L．因此，改进的GA法具有很好的全局优化性能，不但可以有效克服传统遗传算法所存在的缺陷，而且求解过程简单，优化结果合理，在水污染控制规划中具有良好的应用前景．

Application of Improved GA Method in Multi-objective Programming for Water Pollution Control Systems

This paper used an improved Genetic Algorithm method for the multi-objective programming of water pollution control systems. The algorithm adopted the decimal system coding theory, variable weight comprehensive distance evaluation model and multiple-chromosomes crossover mechanism. An example was cited to show that the improved algorithm was able to overcome the disadvantages of the Simplex Genetic Algorithm (SGA) and facilitate the solving process and get satisfactory optimal results. The optimum of adaptive values was between 0.046 8 and 0.056 7. When the variables were 0.033 7,0.566 3,0.4,0,0,0.4,0.137 3,0,0.5,0,0.3,0,0.215 4 and 0.134 6, the function would be optimal. The project cost was RMB 3.127 85 million, underground water consumption was 0.301 25 m~3/s and the values of river oxygen dissolving shortage was 6.694 26 mg/L. The improved algorithm have broad prospects in water pollution control systems.