Finding Near-Optimal Bayesian Experimental Designs via Genetic Algorithms
- 1 August 2001
- journal article
- Published by Taylor & Francis in The American Statistician
- Vol. 55 (3) , 175-181
- https://doi.org/10.1198/000313001317098121
Abstract
This article shows how a genetic algorithm can be used to find near-optimal Bayesia nexperimental designs for regression models. The design criterion considered is the expected Shannon information gain of the posterior distribution obtained from performing a given experiment compared with the prior distribution. Genetic algorithms are described and then applied to experimental design. The methodology is then illustrated with a wide range of examples: linear and nonlinear regression, single and multiple factors, and normal and Bernoulli distributed experimental data.Keywords
This publication has 30 references indexed in Scilit:
- Miscellanea. A note on Bayesian design for the normal linear model with unknown error varianceBiometrika, 2000
- Bayesian D-optimal designs on a fixed number of design points for heteroscedastic polynomial modelsBiometrika, 1998
- Optimal two-point designs for the michaelis-menten model with heteroscedastic errorsCommunications in Statistics - Theory and Methods, 1998
- Optimal designs for rational modelsThe Annals of Statistics, 1996
- Optimal Bayesian design applied to logistic regression experimentsJournal of Statistical Planning and Inference, 1989
- Accurate Approximations for Posterior Moments and Marginal DensitiesJournal of the American Statistical Association, 1986
- A note on Bayes designs for inference using a hierarchical linear modelBiometrika, 1980
- A Remark on the Optimal Regression Designs with Previous Observations of Covey-Crump and SilveyBiometrika, 1971
- Uncertainty, Information, and Sequential ExperimentsThe Annals of Mathematical Statistics, 1962
- Application of a Measure of Information to the Design and Comparison of Regression ExperimentsThe Annals of Mathematical Statistics, 1959