View graph of relations

The core of the proposed project is to investigate and further advance multi-objective, multi-reservoir management by use of surrogate modelling with evolutionary algorithms within a simulation-optimisation framework for optimisation of water resources systems under different climate projections. The research will focus on development of downscaling techniques and weather generators, as well as multi-fidelity and adaptive surrogate modelling techniques that use a sequence of surrogates with varying fidelity in the optimisation. Surrogate models are continuously updated by models with greater physical fidelity as the optimisation proceeds. The developed techniques will be tailored towards application in complex water resources systems, involving multiple objectives as well as multiple reservoir systems. Special emphasis will be given to develop cost-effective methods that are feasible for real-time applications, where real-time data and forecast information are used in the optimisation process. An important aspect that will be considered in the project is how to handle uncertainties and stochasticity in the optimisation process. Uncertainties are related to the inherent errors in the input data used to force the modelling system as well as errors in the modelling system itself. Stochasticity is related to the inherent stochastic nature of the meteorological variables forcing the modelling systems. Uncertainties and stochasticity are normally represented by multiple system states in an ensemble setting, which adds an additional computational challenge to the optimisation problem. The use of weather generators and surrogate modelling is seen as a viable way to solve this problem. The developed techniques will be tested and evaluated on the Red River basin reservoir system in Vietnam.
StatusCurrent
Period01/01/0930/06/12
Financing sourceUnknown
Research programmeUkendt
Amount0.00 Danish kroner
Download as:
Download as PDF
Select render style:
ShortLong
PDF
Download as HTML
Select render style:
ShortLong
HTML
Download as Word
Select render style:
ShortLong
Word

ID: 2283807