Application of multiobjective optimization methods for urban water management: a case study for Canberra water supply system

Planning and management of urban water supply headworks systems is a complex task which gives rise to a large number of infrastructure and operating policy options. This is often complicated by the existence of multiple competing objectives, whereby a gain in one particular objective may result in a loss in another objective. The solution to these problems requires simultaneous consideration of conflicting objectives. Multiobjective optimization deals with the process of simultaneously optimizing two or more conflicting objectives. In this study, the performance of two optimization methods, the ε-dominance genetic algorithm and ant colony optimization algorithm, is compared for the Canberra water supply headworks system. The case study employed two objectives, minimizing present worth cost and time spent in restrictions, and 13 decision variables associated with infrastructure and operating variables. While both methods produced comparable Pareto fronts, the genetic algorithm consistently produced better results and converged faster. This case study demonstrates the potential of multi-criterion optimization to solve realistic problems, thus freeing the systems planner from tedious and most likely inefficient exploration of the solution space