Keywords
global optimization, evolutionary and genetic algorithms, adaptive cluster covering, models calibration
Start Date
1-7-2006 12:00 AM
Abstract
Modelling and decision making related to environmental problems need adequate optimization methods and tools. In case the objective function to be minimized is not known analytically and no assumption can be made about the number of its extrema, gradient-based methods are inapplicable and direct multi-extremum (global) methods must be used. Apart from the popular evolutionary and genetic algorithms, other methods appear to be at least as effective and efficient. Nine algorithms were implemented in the GLOBE global optimization system, and they are compared in terms of effectiveness (accuracy), efficiency and reliability on several benchmark and hydrologic modelling problems.
Experiences in using evolutionary and non-evolutionary optimization methods in models calibration
Modelling and decision making related to environmental problems need adequate optimization methods and tools. In case the objective function to be minimized is not known analytically and no assumption can be made about the number of its extrema, gradient-based methods are inapplicable and direct multi-extremum (global) methods must be used. Apart from the popular evolutionary and genetic algorithms, other methods appear to be at least as effective and efficient. Nine algorithms were implemented in the GLOBE global optimization system, and they are compared in terms of effectiveness (accuracy), efficiency and reliability on several benchmark and hydrologic modelling problems.
