Integer Coding Genetic Algorithm for Optimization of Tree-Type Pipe Network Based on the Graph Theory

An optimization model for tree-type pipe network is established, in which the minimal investment is taken as objective function. The improved integer coding genetic algorithm is designed to optimize the model, through the coding mode design, the optimizations of pipe network layout and pipe diameter are achieved simultaneously, and the pipe diameter restriction is satisfied automatically. In the algorithm operational process, based on the similarity between tree-type pipe network and spanning tree in the graph theory, in the operations of initial solutions creating, crossover and mutation, the method based on loop is adopted, the number of infeasible solutions is avoided or reduced, so the computational efficiency of the method is enhanced.