Multi-criteria decision-making methods for optimal design of intermittent water distribution network
Katineni Ramani,Umamahesh V. Nanduri
DOI: https://doi.org/10.2166/ws.2024.046
2024-03-06
Water Science & Technology Water Supply
Abstract:Multi-objective optimization of water distribution networks (WDNs) is crucial for achieving a balance between cost, equity, and reliability. The present study conducts a comparative analysis, evaluating the effectiveness of four multi-objective optimization algorithms implemented within the EPANET framework for the optimal design of three distinct WDNs taken from the literature. The networks were optimized considering three objectives: maximizing network resilience, and uniformity coefficient (CU, a measure of uniformity), alongside minimizing the cost of the network. Further, multi-criteria decision-making (MCDM) analysis is carried out, encompassing various MCDM techniques and weighting methods, to explore the sensitivity of solution rankings. In Network 1, non-dominated sorting genetic algorithm 2 (NSGA-II) produced suboptimal solutions with CU and In values below 0.4, while caRamel, multi-objective particle swarm optimization based on crowding distance (MOPSO-CD), and multi-objective evolutionary algorithm with decomposition (MOEA/D) showcased increasing CU and In values towards 1, paralleled by a rise in network cost. The performance of NSGA-II in Network 3 was unsatisfactory (CU 0.98) for reliability and (>0.5) for CU. The findings offer valuable insights into cost-effective strategies for achieving equity in water supply without substantial impact on overall network cost. The findings of the study also underscore the sensitivity of ranking outcomes in MCDM analysis to the choice of technique and weighting method.
environmental sciences,water resources,engineering, environmental