Genetic algorithms for design of pipe network systems

Abstract

In the last three decades, a significant number of methods for optimal design of pipe network systems have been developed using linear programming, non-linear programming, dynamic programming, enumeration techniques, and genetic algorithm (GA). This paper presents a genetic algorithm (GA) approach to the design of a pipe network systems. The objectives considered are minimization of the network cost in the practical requirement. Of all the preceding methods, GA based methods appear to be robust, as they can handle discrete pipes sizes with ease and produce a set of alternatives. By judgmentally selecting the Ruey-Fang district water supply system for the case study and by comparing the data gathered from the case study, this study also aims to verify the efficacy of the proposed method. The finding indicates that the proposed method is superior to enumeration techniques with respect to solution speed and cost.