2008
EMSS 2008 Proceeding
Incorporating physical knowledge about the formation of nitric oxides into evolutionary system identification
Authors: Stephan Winkler, Markus Hirsch, Michael Affenzeller, Luigi del Re, Stefan Wagner
Abstract
Genetic programming (GP) is an evolutionary optimizationmethod that has already been used successfully forsolving data mining problems in the context of severalscientific domains. For example, the identification ofmodels describing the nitric oxides (NOx) emissions ofdiesel engines has been investigated intensively, verypromising results were obtained using GP.In the standard GP process, all model structures (aswell as parameter settings) of models are created duringan evolutionary process; populations of models areevolved using the genetic operators crossover, mutationand selection. In this paper we discuss several possibilitieshow a priori knowledge can be integrated into theGP process; we have used physical knowledge about theformation of NOx emissions in a BMW diesel engine,test results are given in the empirical tests section.
I3M Scientific Sponsors
I3M Industrial Sponsors
I3M Media Sponsors
