Premium
Using Genetic Algorithms to Rehabilitate Distribution Systems
Author(s) -
Wu Zheng Y.,
Boulos Paul F.,
Orr Chun Hou,
Ro Jun Je
Publication year - 2001
Publication title -
journal ‐ american water works association
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.466
H-Index - 74
eISSN - 1551-8833
pISSN - 0003-150X
DOI - 10.1002/j.1551-8833.2001.tb09335.x
Subject(s) - sizing , genetic algorithm , solver , pipe network analysis , computer science , selection (genetic algorithm) , range (aeronautics) , mathematical optimization , operations research , engineering , machine learning , art , physics , mathematics , visual arts , thermodynamics , programming language , aerospace engineering
Many US water supply infrastructure systems are nearly 100 years old. Where as some are working nearly as efficiently as when they were first installed, others are showing their age with clogging, leaks, breaks, and service disruptions. To pinpoint the best solutions for each particular project, past practices often involved a tedious trial‐and‐evaluation procedure that seldom led to the most effective or most economical solutions for upgrading pipe networks. Because these upgrades and modifications are usually very expensive, it is imperative that sound analytical procedures are used to select economical design and rehabilitation alternatives. The fast messy genetic algorithm (fmGA), based on outcomes seen in nature, automatically determines low‐cost design rehabilitation and expansion alternatives that best meet prescribed system performance criteria. This methodology helps water utilities streamline their network rehabilitation and capital expenditure requirements and determine facility sizing and timing for system enhancements and expansions. This optimization procedure is a variation of the GA, a popular intelligent search procedure based on the mechanics of natural selection and evolution. The hydraulic simulation is performed using an extended version of the EPANET water distribution network solver. It selects least‐cost improvement alternatives that satisfy designated constraints on the network hydraulic performance for any given range of demand loading and operating conditions. This research compared outcomes using the fmGA method with three water distribution system examples taken from the literature. The fmGA found cost‐effective solutions much more quickly than the previously used analytical methods.