Genetic Algorithm Optimization for Water Supply Systems Planning

Abstract

Water supply systems are built ahead of time to meet future demand. Investments are made in steps to cover needs up to a planning horizon. The objective of long term planning optimization is to come up with the least cost scenario for investment decisions. Potential savings could be very significant and influence the decision to invest or not for a given project. These also allow the choice of a feasible and optimal alternative amongst several project scenarios.

The present paper describes new approaches to long-term planning optimization using genetic algorithm search methods. The new approaches consist in determining the best phasing for longterm water projects. The formulation of the optimization problem is described along with three different genetic algorithm approaches. A case study based on the long-term water supply requirements of a region in North America is investigated. Finally, the results for the different GA formulations are presented and discussed.