树状灌溉管网优化的整数编码改进遗传算法

Integer Coding Improved Genetic Algorithm for Optimization of Tree-type Irrigation Pipe Network

建立了以投资最小为目标函数的树状灌溉管网优化数学模型,并采用整数编码改进遗传算法对模型进行求解。针对完整树状管网优化的特点,通过对整数编码的编码方式进行改进,实现了同时对管网中管线连接关系和管径的优化。改进了目标函数映射到适应函数的变换方式,设计了随机多父辈交叉方法,提出了2种与优化问题相适应的变异方法,从而提高了算法的优化性能。

An optimization model for tree-type irrigation pipe network was established,the minimal investment was taken as objective function,and integer coding improved genetic algorithm was adopted for solving the model. Aiming at the features of the complete optimization for tree-type pipe network,through the integer coding mode design,the optimizations of pipe line connected relationship and pipe diameter were realized simultaneously. Mapping mode of the fitness function was improved,random multi-parent crossover was designed,and two kinds of mutation methods compliant with the problem were proposed,so the optimal performance of the algorithm was enhanced.