Using Genetic Programming in Industrial Statistical Model Building

Abstract

The chapter summarises the practical experience ofintegrating genetic programming and statisticalmodelling at The Dow Chemical Company. A uniquemethodology for using Genetic Programming instatistical modeling of designed and undesigned data isdescribed and illustrated with successful industrialapplications. As a result of the synergistic efforts,the building technique has been improved and the modeldevelopment cost and time can be significantly reduced.In case of designed data Genetic Programming reducedcosts by suggesting transformations as an alternativeto doing additional experimentation. In case ofundesigned data Genetic Programming was instrumental inreducing the model building costs by providingalternative models for consideration.