1st International Congress on Environmental Modelling and Software - Lugano, Switzerland - June 2002
Keywords
optimization, multi-objective, pareto solutions, water distribution system, design
Start Date
1-7-2002 12:00 AM
Abstract
In many optimisation problems, analysts are often confronted with multiobjective decision problems. The most common purpose of an analysis is to choose the best trade-offs among all the defined and conflicting objectives. However, many optimisation studies are formulated as a problem whose goal is to find the “best” solution, which corresponds to the minimum or maximum value of a single objective function that lumps all different objectives into one. Water distribution system design is a multiobjective problem for which it is difficult to identify the true benefits and constraints due primarily to the uncertainty in future demands. This paper shows some shortcomings of the use of single-objective optimisation for water distribution system design and introduces a genetic algorithm multiobjective model that promises to ease the difficulties in applying optimisation and providing decision support for that important problem. The optimisation model used in this paper utilises simple and intuitive objectives and constraints that are not difficult to formulate in mathematical terms. Those objectives allow a decision-maker to visualise the trade-offs between different benefits and costs, and more importantly to consider uncertainty in future demands and performance levels. This type of optimisation could also take into account that the system needs to be implemented in stages.
Single-objective vs. Multiobjective Optimisation for Integrated Decision Support
In many optimisation problems, analysts are often confronted with multiobjective decision problems. The most common purpose of an analysis is to choose the best trade-offs among all the defined and conflicting objectives. However, many optimisation studies are formulated as a problem whose goal is to find the “best” solution, which corresponds to the minimum or maximum value of a single objective function that lumps all different objectives into one. Water distribution system design is a multiobjective problem for which it is difficult to identify the true benefits and constraints due primarily to the uncertainty in future demands. This paper shows some shortcomings of the use of single-objective optimisation for water distribution system design and introduces a genetic algorithm multiobjective model that promises to ease the difficulties in applying optimisation and providing decision support for that important problem. The optimisation model used in this paper utilises simple and intuitive objectives and constraints that are not difficult to formulate in mathematical terms. Those objectives allow a decision-maker to visualise the trade-offs between different benefits and costs, and more importantly to consider uncertainty in future demands and performance levels. This type of optimisation could also take into account that the system needs to be implemented in stages.
