Global Maximum Entropy Minimum Cost Design of Water Distribution Systems

摘要

The problem of designing water distribution systems to deliver the Global Maximum Entropy (GME) flows was hitherto hampered by the multiplicity of feasible flow directions associated with looped water distribution networks. This paper addresses this issue by presenting a new multi-objective approach to the design optimization of water distribution systems based on a robust and fast genetic algorithm, namely NSGA-II. The decision variables used in the approach are standard discrete pipe diameters. By integrating a hydraulic simulator with an algorithm for detection of flow directions, the approach is capable of achieving the Global Maximum Entropy Minimum Cost (GMEMC) design for a fixed layout. To ensure that the achieved design is near global, the approach does not assign penalties to any generated design. The new approach is demonstrated by designing a well known hypothetical network in the literature.