基于改进遗传算法的水库群防洪优化调度

为解决标准遗传算法(SGA)在处理高维复杂问题中寻优能力不足、计算时间过长、易早熟的问题,首先改进映射规则,仅选择部分决策变量进行编码,采取新的映射规则将其解码映射至解空间,实现高维问题的降维;其次使用针对不同约束条件的自适应罚函数对算法收敛性进行改进;最后以北方某流域并联水库防洪调度为例,建立防洪优化调度模型,并运用改进遗传算法进行求解。结果表明,该方法在解决高维复杂水库群防洪优化调度问题时具有求解速度快、寻优结果好的优点,有一定的实用性。

Optimal Operation of Flood Control for Reservoir Group Based on Improved Genetic Algorithm

In order to overcome some defects caused by standard genetic algorithm (SGA) to deal with high-dimensional complex problems, such as poor optimization ability, long-time computation and premature, an improved GA was proposed in this paper. First of all, encoding a part of decision variables and decoding them to the solution space with new mapping rules was put forward to reduce the dimension of high-dimensional problems. Then, the convergence property of SGA was modified by using adaptive penalty function to handle various constraints. Taking the flood control operation of northern parallel reservoirs as an example, the model of flood control operation of the reservoir was built and solved by utilizing improved genetic algorithm. The result shows that this method has advantages of quick convergence speed and getting better result when coping with high-dimensional complicated flood control operation of reservoir group. Thus, the method is practical to some extent.