We describe in this paper an extension of standard Genetic Programming where the terminal set of the algorithm is expanded with a set of basic models generated offline using a deterministic approach. The new algorithm called EMM-GP (Genetic Programming based Evolvement of Models of Models) uses a specialized mutation operator to sample this set during the recombination phase in order to create models that reuse valuable building blocks. In this work, a preprocessing of model set by means of clustering is considered.
|Journal||IOP Conference Series: Materials Science and Engineering|
|Publication status||Published - 29 Jan 2020|
|Event||2nd International Scientific Conference on Advanced Technologies in Aerospace, Mechanical and Automation Engineering, MIST: Aerospace 2019 - Krasnoyarsk, Russian Federation|
Duration: 18 Nov 2019 → 21 Nov 2019