SaDE algorithm for reliability-based design of water distribution networks

Design of Water Distribution Networks (WDNs) is an NP hard problem, which is typically an optimization problem consisting of a large search space even for a small sized WDN. Design of a robust WDN requires consideration of aspects such as reliability, resiliency etc., which further increases the complexity of the problem. This paper presents an efficient way of reliability-baseddesign of WDNs using Self-adaptive Differential Evolution (SaDE) algorithm, which is an evolutionary optimization technique and animprovised version of the DE algorithm. The reliability of the WDNs is estimated as the level of demand satisfaction considering the mechanical failure of pipes. Simulations are performed using EPANET for generating the pressure and flow values using demand driven analysis, for various failure and working conditions. The efficacy of the proposed approach is tested by applyingitto two WDN problems. A thorough comparison is made between the performance of SaDE and DE algorithms. It is found that the SaDE converges much faster than DE, with high success rate, thus making it a more preferable option for solving WDN design problems and for reliability-based design of WDNs.