Application of an evolutionary algorithm to the inverse parameter estimation of an individual-based model

摘要: Abstract Inverse parameter estimation of individual-based models (IBMs) is a research area which still in its infancy, context where conventional statistical methods are not well suited to confront this type with data. In paper, we propose an original evolutionary algorithm designed for the calibration complex IBMs, i.e. characterized by high stochasticity, uncertainty and numerous non-linear interactions between parameters model output. Our corresponds variant population-based incremental learning (PBIL) genetic algorithm, specific “optimal individual” operator. The method presented detail applied OSMOSE. performance evaluated estimated compared independent manual calibration. results show that automated convergent inverse significant improvement existing ad hoc IBMs.