A Novel Self-Adaptive, Non-Metaphor-Based FISANET Framework for Pressure Dependent Optimization of Water Distribution Networks

Water distribution networks (WDNs) are a critical infrastructure that connects sources of water to end consumers, where cost optimization remains a fundamental challenge due to substantial construction investments required. This research introduces FISANET, a novel self-adapted, metaphor free optimization approach for cost-effective water distribution network design. This FISANET approach is developed in a Python environment with integration in EPANET 2.2 hydraulic solver, which eliminates the need for complex parameter tuning while maintaining computational efficiency. Performance evaluation is being conducted on three established benchmark networks, namely the Two-Loop Network (TLN), the Hanoi Network (HN), and the New York Tunnel (NYTN) Network, along with a real-world field water distribution network of the School of Planning and Architecture (SPAN), Bhopal, M.P., India. Results demonstrate that FISANET achieves optimized cost solutions with significantly reduced function evaluations compared to other existing approaches. Also, this pressure-driven demand analysis (PDD) integrated FISANET approach consistently outperformed traditional demand-driven Analysis (DDA) approaches, achieving a cost reduction to 1.62% and found new best solution for the New York tunnel network, while also achieving best-known solutions for Two-loop and Hanoi networks. The self-adaptive nature of FISANET requires less computational costs associated with algorithm calibration, making it particularly suitable for practical engineering applications where rapid convergence and reliability are essential for infrastructure design decisions.