Cooperative Random Walk for Pipe Network Layout Optimization

Pipe network layout optimization is an important and difficult problem in pipe network engineering. Differing from the previously proposed approaches for various optimization formulations, the authors use a methodology for pipe network optimization consisting of network layout optimization coupled with pipe sizing analysis. The methodology is based on a novel heuristic optimization algorithm inspired by Ant Colony Optimization, called Cooperative Random Walk (CRW). CRW successfully finds an optimal network layout for a range of pipe diameter values, as shown on a test base network of 40 nodes and 68 pipes with one source and 12 consumer nodes. The results clearly show how larger pipe diameters allow for simpler network layouts, while smaller pipe sizes automatically imply more complex and longer networks. Although CRW is tested on a water supply network problem, it is suitable for other network optimizations as well.