Increasing the capacity of water distribution networks using fitness function transformation.

Even the most carefully designed water distribution network (WDN) can suffer from local capacity deficiencies as a result of the quick and unpredictable growth of the urbanization of new industrial sites. To solve this problem, this paper focuses on the identification of the best possible location for a new pipeline within an existing WDN, which maximizes the node-wise capacity. To determine the optimal solution, a parameter, namely pressure sensitivity, is defined, which can localize nodes with local capacity problems computationally efficiently. During our research, a fitness function transformation technique was defined, which increases the effectivity of the optimization on a larger scale by the formulation of a feasible fitness function. Combining this technique with an extended version of the genetic algorithm, the topology of eleven real-life WDN was optimized. A scrutiny is performed on three networks, highlighting typical deficiencies.