Document Type
Presentation
Publication Date
8-1-2014
Abstract
Management of a water supply system from the point of view of the conjunctive use of water sources is a very complex problem whose solution is not just obtained using analytical models but also through a negotiation process among stakeholders and in which Public Bodies have a main role. For these reasons, this problem has been addressed using conservative approaches based on simulation models or simulation – linear optimization models parameterized using few parameters. These approaches have certain strengths but also certain drawbacks or constraints. In general these conservative approaches are already covered by existing generalized modelling tools (i.e. Aquator, Oasis and Riverware) that may be used to estimate a solution through a longer or shorter trial and error process. However, these conservative approaches and the corresponding generalized modelling tools have drawbacks and constraints when dealing with certain complexities of water supply systems (i.e.: non-linearity, uncertainty or stochastic nature) that may prevent them of finding an optimal solution. Therefore, a functionality that current modelling systems are starting to offer to overcome this drawback is the capacity to be linked with external software modules mainly implementing accurate and complex optimization methods (i.e.: Evolutionary Algorithms / SolveXL tool). This paper identifies and tests suitable Simulation-Optimization approaches found in existing Generalized Modeling Tools for optimizing operating rules of multiple source water supply systems in terms of system reliability and resulting operating costs: a special focus on Genetic Algorithms is considered. The main purpose is to find out whether these approaches are already covering the decision support needs of managers, Public Bodies or other stakeholders involved in the operation of these systems, or ‘ad-hoc’ tools may be needed. This process leads to identify strengths and weaknesses of these modeling approaches found in existing Generalized Modeling Tools.
Comments
Session R63, Parameter Estimation: Simulation Optimization