Differential evolution algorithm for optimal design of water distribution networks

Abstract

Water distribution networks are considered as the most important entity in the urban infrastructure system and need huge investment for construction. The inherent problem associated with cost optimisation in the design of water distribution networks is due to the nonlinear relationship between flow and head loss and availability of the discrete nature of pipe sizes. In the last few decades, many researchers focused on several stochastic methods of optimisation algorithms. The present paper is focused on the Differential Evolution algorithm (henceforth referred to as DE) and utilises a similar concept as the genetic algorithm to achieve a goal of optimisation of the specified objective function. A simulation–optimisation model is developed in which the optimization is done by DE. Four well-known benchmark networks were taken for application of the DE algorithm to optimise pipe size and rehabilitation of the water distribution network. The findings of the present study reveal that DE is a good alternative to the genetic algorithm and other heuristic approaches for optimal sizing of water distribution pipes.