Model selection using a simplex reproduction genetic algorithm
Signal Processing
Key concepts in model selection: performance and generalizability
Journal of Mathematical Psychology
Discovering the Structure of Partial Differential Equations from Example Behaviour
ICML '00 Proceedings of the Seventeenth International Conference on Machine Learning
An Indirect Block-Oriented Representation for Genetic Programming
EuroGP '01 Proceedings of the 4th European Conference on Genetic Programming
Introduction to Evolutionary Computing
Introduction to Evolutionary Computing
The Knowledge Engineering Review
Ecological Model Selection via Evolutionary Computation and Information Theory
Genetic Programming and Evolvable Machines
Effects of Diversity on Optimality in GA
AICI '09 Proceedings of the International Conference on Artificial Intelligence and Computational Intelligence
Hi-index | 0.00 |
The addition of two emerging technologies (evolutionary computation and ecoinformatics) to computational ecology can advance our ability to build better ecological models and thus deepen our understanding of the mechanistic complexity of ecological systems. This article describes one feasible approach toward this goal-the combining of inductive and deductive modeling techniques with the optimizing power of simple algorithms of Darwinian evolution that include information-theoretic model selection methods. Specifically, the author shows a way to extend classic genetic algorithms beyond typical parameter fitting of a single, previously chosen model to a more flexible technique that can work with a suite of possible models. Inclusion of the Akaike information-theoretic model selection method within an evolutionary algorithm makes it possible to accomplish simultaneous parameter fitting and parsimonious model selection.Experiments with synthetic data show the feasibility of this approach, and experiments with time-series field data of the zebra mussel invasion of Lake Champlain (United States) result in a model of the invasion dynamics that is consistent with the known hydrodynamic features of the lake and the motile life history stage of this invasive species.The author also describes a way to extend this approach with a modified genetic programming algorithm.