Improved Genetic Algorithm for Optimization of Multi-Source Pumping Tree-Type Pipe Network

An optimization model for multi-source pumping tree-type pipe network is established, in which the investment and operating cost are considered simultaneously. Useing graph theory knowledge, the essence of pipe layout optimization for multi-sources tree-type pipe network is searching for a group of optimal subtrees. In the complete optimization of pipe network, the pipe layout mode and pipe diameter are needed to optimize simultaneously, so the dual coding genetic algorithm combining binary coding with integer coding is adopted, and the pipe diameter restriction is satisfied automatically. According to the tree character, method based on cycle is adopted in the operational process of binary coding, the partheno-genetic operation is designed to produce new individual, the infeasible solutions are avoided. Example shows the efficiency of the algorithm.