Multiobjective optimization of water distribution system under uncertainty using robust optimization

Abstract

[EN] Uncertainty is inevitable while trying to tackle any real-world problem, Water distribution system design problem is not an exception for that. Among the many uncertainties involved in the design problem, demand uncertainty is most important. The current study aims to provide efficient designs that can handle the predicted variations in demand without compromising the resilience of the network system. Most previous studies explored stochastic solutions to handle the uncertainty that consumed immense computational power. Innovations in developing non-probabilistic techniques like robust optimization paved a way to reduce computational time and handle uncertainty efficiently. A new methodology is proposed in this study to obtain efficient designs for the multiobjective design problem of WDS under uncertainty. The proposed methodology uses a combination of robust optimization approaches to handle the uncertainty and a multiobjective cuckoo search algorithm. The proposed methodology is applied to a standard benchmark water distribution system problem, and the designs are compared with the nominal designs obtained when demand is assumed as certain. Furthermore, the effect of considering different uncertainty sets are discussed.