Font Size: a A A

Research On The Application Of Improved Hybrid Genetic Algorithm To Optimal Design Of Water Distribution System

Posted on:2007-12-30Degree:DoctorType:Dissertation
Country:ChinaCandidate:C S ChuFull Text:PDF
GTID:1102360212470857Subject:Environmental Engineering
Abstract/Summary:PDF Full Text Request
Water supply system is the important public facility for the city's construction and development, and it is also the indispensable substantial element for people. However, water distribution system (WDS) plays an important part of the water supply system, it not only accounts for 50-80% cost of the total invest, but also has great influence on the power expenditure and the reliability of the whole system. Therefore, it is very important and significant to make optimal design for the WDS.By analyzing the existed optimal mode of design to WDS, the author presents a multi-objective model based on economy and reliability. In the process of setting up the model, the reliability of WDS is defined as"weighted mean of nodal extra pressure"and"pipe network resilience", and which will make it easy and direct to evaluate the reliability of WDS; at the same time, the annual cost is looked as the economical characteristic of WDS.Aiming at model's essence of combinational optimization with discrete variable, the author has a research on the optimal methods such as linear programming, nonlinear programming, dynamic programming, artifical neural network method, genetic algorithm. By analyzing the shortcomings of these theories and methods, the author puts forward an optimal method named improved hybrid genetic algorithm (IHGA) to solve the optimal mode. Genetic algorithm (GA) is a kind of stochastic searching algorithm that simulates the mechanism of natural selection and nature heredity. Furthermore, GA has the characteristic of expandability and is easily to combine with other technique to form hybrid genetic algorithm (IGA) that is usable for complicated optimal calculation.To overcome the faults of GA, the paper introduces the migrate strategy to improve GA and makes it possible to get optimal result by less calculation; the integer-coding technique is used to pipe diameter to avoid rounding; and the penalty strategy is adopted to the violating pipe to eliminate the infeasible solution.In order to get pipe flow and node pressure, it is necessary to have hydraulic analysis before the calculation of object function and fitness for the optimal design.
Keywords/Search Tags:optimal design of water distribution system, multi-objective model, reliability of water distribution system, pipe network resilience, improved hybrid genetic algorithm, migrate strategy, trust region method
PDF Full Text Request
Related items