Research On Improved Shuffled Frog Leaping Algorithm For Upgrading And Expansion Of Urban Water Network

Urban water-supply pipe network system is the lifeline of the city, and also is abridge connecting users and waterworks. However, as work experience increases, waterdistribution system appeared severely corrosive, explosive hose frequent, secondarypollution and problems such as inadequate water supplies, thus making pipeline systemexpansion is necessary and urgent. Scientific pipe extension model is not only affectwater needs of the inhabitants, also affect urban water-supply pipe network systemconstruction and operation and maintenance cost. Based on previous research, thisarticle using novel shuffled frog leaping algorithm to optimize the system and extensionof water supply network.（1）As a new swarm intelligent organic search optimization algorithm, shuffledfrog leaping algorithm (SFLA) combines the advantages of Particle SwarmOptimization algorithm and the merits of memetic algorithms, can develop the ability tosustainable local development capability and global optimization of balance, withsimple concepts, less arguments and good robustness. Considering poor random localsearch, volatile step size defects of SFLA, respectively, this article proposed measuresto improve SFLA to make it perfect.（2）Accurate prediction of medium and long term water demand of the city ispremise and basis of rebuilding in urban water-supply pipe network system optimization.Using improved SFLA to optimize the connection weights and thresholds of BP neuralnetwork, and to build SFLA-BP network to forecast urban water demand formedium-and long-term. Case study of some city, the paper used principal componentanalysis to obtain the factors affecting water demand, and then used SFLA-BP networkto forecast water demand, compared to BP and PSO-BP network, SFLA-BP networkcomputing time shorter, forecasting higher accuracy, thus validating the goodcomputing performance and global search capabilities of SFLA.（3）Based on improved SFLA to establish upgrading and expansion model ofurban water network, the paper introduced hydraulic reliability and entropy valuereliability to the objective function except for the economy, while taking advantage ofobjective method of multiplication and division to turn the multi-target objectivefunctions into a single objective function. Using MATLAB as platform, X city forengineering instance, with traditional model as controlled, the paper calculated two models in most adverse points, and hydraulic Fire and accident conditions, and thencompared the two models of economic indicators, reliability indicators and compositeindicators, the results show that, compared with the traditional model:①Regarding average annual power cost and year commuted value of new watertreatment plant, the optimization model cost a little higher, but pipeline builtmaintenance costs were significantly lower, overall, optimization model saved about6%of the cost savings compared to traditional model.②The entropy reliability of optimization model was similar to traditional model,but with better reliability of hydraulic.③As for evaluation index: Required cost per reliability, the result of optimizationmodel was lower, verified that the upgrading and expansion model of urban waternetwork based on improved SFLA was feasible and convenient in the practicalapplication of engineering.