The Research Of Optimal Design Of Water Distribution Network Based On Hybrid Particle Swarm Genetic Algorithm

Water distribution system is an important part of the urban water supply system, generally, its investment accounts for the 50%-80% of the total investment to water distribution system. Water supply network system is composed of a variety of facilities for water distribution system, which directly affects the running costs of urban water supply systems and water supply reliability. With the expansion of urban scale, water supply network has been developed toward large-scale and complex direction. Therefore, optimal design of water distribution system plays an important and meaningful role.At present, the use of genetic algorithms for pipe network optimization problem is gradually being valued by domestic and foreign researchers. Based on the current research results of optimal design of water distribution system, the basic principle and implement method of genetic algorithm, this paper, on the one hand, thinking that mutation operator of genetic algorithms is difficult to determine and post-convergence slow is so easy to fall into local optimization problem, genetic algorithm (GA) and particle swarm optimization (PSO) are used conjointly. Using particle swarm optimization algorithm to construct the mutation operator of genetic algorithm--PSO mutation operator, in this way hybrid particle swarm-genetic algorithm (PSOGA) can be formed. On the other hand, on the basis of economical object function of water distribution system, weighted mean of nodal extra pressure and reliable object function of pipe network resilience are added to constitute multi-objective mathematical model of water distribution system.In the application of hybrid particle swarm-genetic algorithm to solve multi-objective mathematical model of water distribution system, diameter variables are encoded by integer coding technology to avoid other method raises second roundness diameter. Meanwhile, penalty function is used to improve the object function, and penalty technology is used to process barrels that violate economy velocity constraints and nodes that violate nodes pressure constraints, and only in this way the infeasible solution can be eliminated.Finally, the theories and methods of hybrid particle swarm-genetic algorithm and optimization model are tested by studying a typical representative of optimal design of water distribution system case, and calculated results and the results of basic genetic algorithm are compared from the aspect of economy, hydraulic performance and reliability to fully explain the important theoretical and practical value of hybrid particle swarm-genetic algorithm in the design of water distribution system.