ROBUST DESIGN AND MANAGEMENT OF WATER SYSTEMS: HOW TO COPE WITH RISK AND UNCERTAINTY?

摘要

The implications of uncertainty in input data and model parameters are not always understood or considered when designing urban water systems. There is an obvious need to develop design methods that can model uncertainties and produce 'robust' designs. These designs should provide adequate service to customers despite fluctuations in some or all of the design parameters. This paper presents a general methodology for robust design of such systems. The methodology is based on closely integrating reduced Monte-Carlo sampling with multiobjective Genetic Algorithms. An application to the New York Tunnels rehabilitation problem illustrates the effectiveness of the methodology. The results show that the robust design methodology presented here seems to be capable of identifying Pareto optimal fronts for uncertain input variables while achieving significant computational savings when compared to the full MC sampling technique.