Water distribution network optimization using maximum entropy under multiple loading patterns

This paper proposes a maximum entropy-based multi-objective genetic algorithm approach for the design optimization of water distribution networks (WDNs). The novelty is that in contrast to previous research involving statistical entropy the algorithm can handle multiple operating conditions. We used NSGA II and EPANET 2 and wrote a subroutine that calculates the entropy value for any given WDN configuration. The proposed algorithm is demonstrated by designing a six-loop network that is well known from previous entropy studies. We used statistical entropy to include reliability in the design optimization procedure in a computationally efficient way.