遗传算法是通过对样本中个体的不断改进来寻找各类问题的最优解。由于标准遗传算法（SGA）存在收敛性及个体适应度求解方面的困难，在研究中，通过对SGA中遗传算子改进，特别是对选择算子的改进，提出了一种改进遗传算法（AGA），并将它应用于水库优化调度中。改变通常以水位变化序列为基础的遗传算法编码方案，通过数组存储水库库容状态，并以各库容状态对应的数组下标为基础进行遗传算法编码，通过实例，表明AGA对水库优化调度问题具有良好的适应性，同时结合数组存储理论的遗传算法编码方法简化了水库优化调度遗传算法的实现过程。

Genetic algorithms search for the optimal solution by continually improving the individual of the population. Because of the difficulty in convergence and solving of individual fitness, standard genetic algorithm (SGA) is not used widely. Based on the improvement of SGA, especially the improvement of the selection operator in SGA, a new genetic algorithm(AGA) is proposed to solve the problems about the optimal operation of reservoirs. A new coding method is presented which is based on the subscript sequence of reservoir capacity array other than the water level sequence. An engineering example illustrates that AGA is much more efficient than SGA, and also the new coding method predigests the course of genetic algorithm in the optimal operation of reservoirs.