The Optimal Design Of Urban Water Distribution Systems Via Improved Particle Swarm Optimization

Urban water distribution network system is an important infrastructure of urbanconstruction and industrial production. With the expansion of the city scale, the size ofpipe network system is gradually showing a complex, large-scale trend and fund inputfor construction of pipeline systems increases as well. Generally, almost5%-10%ofthe project investment can be saved by optimizing calculation as investment for waterdistribution network accounts for three-quarters of the total investment. Due to itsgreat economic and practical significance, how to scientific optimization design ofwater supply pipe network has become the concern of the domestic and foreignexperts.The mechanism of the Particle Swarm Optimization (PSO) algorithm has somefeatures: the simplicity, less adjustable parameters, easy to implement, strong globalconvergence ability and not rely on the characteristic information of specific problem.Therefore, in this paper the problem of water distribution network are optimized bythe PSO. The main works are as follows:1. Considering that PSO algorithm has the problem of easily getting in the localminimum and therefore getting difficult in finding the optimal solution when it is usedto optimize water distribution networks. Based on the analysis of inertia weightimpact on the performance of PSO algorithm, this paper proposes a dynamicadjustment strategy of inertia weight based on the particle concentration information.The inertia weight can be adaptively adjusted according to the distribution of particlesin the population in the course of evolution, and improved algorithm can fully balancethe global exploring ability in the process of optimization algorithm and localdevelopment ability, though which the convergence precision of the algorithm can beimproved. The improved Particle Swarm Optimization is tested on four benchmarkfunctions and the simulation results show that comparing with the traditional strategyof inertia weight, the improved strategy is more able to adapt to the performance ofthe particle dynamic search.2. In order to solve highly nonlinear, discrete combination optimization problemswith constraint conditions like optimization design of water supply pipe network, amodified particle swarm hybrid optimization algorithm is put forward. Based on the improvement of dynamic adjustment inertia weight, the Extremal Optimization (EO)algorithm was introduced into the improved PSO algorithm. By using the local searchability of EO, species diversity of algorithm increases and algorithm can effectivelyjump out of local optimum. According to the distribution of the solution of theoptimization design of water supply pipe network, the adaptive penalty functionmethod is used to deal with the constraint conditions of network optimization and toimprove the search efficiency of the algorithm. Application of improved particleswarm hybrid optimization algorithm to solve this problem and the simulation resultsshow that the improved particle swarm hybrid optimization algorithm has fasterconvergence rate and smaller opportunity of trapping in local optimal than normalalgorithm，and obtains the better project cost.3. In this paper, a stand-alone version of the intelligent computing decisionssystem of urban water distribution network is developed in Visual Studio2010andMATLAB environment, and latest intelligent optimization algorithms are embeddedin the system by using database. This system is capable of optimizing the calculationof water supply network, and quickly and accurately obtaining the total cost andoptimal network diameter. Besides, it can also provide optimization decisionsaccording to the actual condition in the stage of laying pipeline. Moreover, data wouldbe compared and analyzed after setting different parameters of operating situation andpipeline cost and an optimum strategy would be offered to decision-makers in anintuitive and.visible way.In general, this paper introduces an improved particle swarm hybrid optimizationalgorithm aimed at solving the existing problems in optimizing water supply pipenetwork, and the results are satisfied without affecting the demand of water supply.Moreover, intelligent computing decisions system of urban water distribution networkis developed, which can offer fundamental data supporting to city planning and designand urban construction.