Linear Combination of Distance Measures for Surrogate Models in Genetic Programming
Surrogate models are a well established approach to reduce the number of expensive function evaluations in continuous optimization. In the context of genetic programming, surrogate modeling still poses a challenge, due to the complex genotype-phenotype relationships. We investigate how different genotypic and phenotypic distance measures can be used to learn Kriging models as surrogates. We compare the measures and suggest to use their linear combination in a kernel.
We test the resulting model in an optimization framework, using symbolic regression problem instances as a benchmark. Our experiments show that the model provides valuable information. Firstly, the model enables an improved optimization performance compared to a model-free algorithm. Furthermore, the model provides information on the contribution of different distance measures. The data indicates that a phenotypic distance measure is important during the early stages of an optimization run when less data is available. In contrast, genotypic measures, such as the tree edit distance, contribute more during the later stages.
KeywordsGenetic programming Surrogate models Distance measures
- 2.Flasch, O.: A modular genetic programming system. Ph.D. thesis, TU Dortmund (2015)Google Scholar
- 5.Parisotto, E., Mohamed, A., Singh, R., Li, L., Zhou, D., Kohli, P.: Neuro-symbolic program synthesis (2016). arXiv e-prints 1611.01855
- 7.Zaefferer, M., Stork, J., Friese, M., Fischbach, A., Naujoks, B., Bartz-Beielstein, T.: Efficient global optimization for combinatorial problems. In: Proceedings of the 2014 Genetic and Evolutionary Computation Conference, GECCO 2014, pp. 871–878. ACM, New York (2014)Google Scholar
- 13.Nguyen, S., Zhang, M., Tan, K.C.: Surrogate-assisted genetic programming with simplified models for automated design of dispatching rules. IEEE Trans. Cybern. 47(9), 1–15 (2016)Google Scholar
- 14.Moraglio, A., Kattan, A.: Geometric surrogate model based optimisation for genetic programming: Initial experiments. Technical report, University of Birmingham (2011)Google Scholar
- 16.Mockus, J., Tiesis, V., Zilinskas, A.: The application of Bayesian methods for seeking the extremum. In: Towards Global Optimization 2, North-Holland, pp. 117–129 (1978)Google Scholar
- 18.Pawlik, M., Augsten, N.: APTED release 0.1.1. GitHub (2016). https://github.com/DatabaseGroup/apted. Accessed 01 June 2017
- 19.Moraglio, A., Poli, R.: Geometric landscape of homologous crossover for syntactic trees. In: 2005 IEEE Congress on Evolutionary Computation, Edinburgh, UK. IEEE (2005)Google Scholar
- 22.Flasch, O., Mersmann, O., Bartz-Beielstein, T., Stork, J., Zaefferer, M.: RGP: R genetic programming framework. R package version 0.4-1 (2014)Google Scholar
- 23.Zaefferer, M.: Combinatorial efficient global optimization in R - CEGO v2.2.0 (2017). https://cran.r-project.org/package=CEGO Accessed 10 Jan 2018