AWWA JAW54061

This study discusses genetic algorithm (GA) optimization, a powerful, practical tool that would assist utilities facing hard decisions about how to meet future growth. GA optimization can be used to identify a range of feasible, low-cost solutions that satisfy specific design and performance criteria. GA analysis fits into the standard distribution system modeling and alternatives evaluation approach and offers the added benefit of minimizing a project's life-cycle costs. This study describes a case study from Grand Prairie, Texas, that demonstrates how one city used GA optimization to identify the best overall water system expansion plan to meet future needs. GA optimization considered such variables as location and size of new pipes and storage; settings of pumps and flow control and pressure-reducing valves; possible elimination of existing control valves, tanks, and pump stations; and, the choice of supply rates from different water source connections. The model optimized not only planned capital improvements (new pipes, tanks, pumps, and valves) but also operational set points and schedules for regulating valves and pumping operations. GA optimization regularly achieves capital cost savings of 20-30%. Includes 12 references, table, figures.