Pareto-Front Exploitation in Symbolic Regression
Symbolic regression via genetic programming (hereafter, referred to simply as symbolic regression) has proven to be a very important tool for industrial empirical modeling (Kotanchek et al., 2003). Two of the primary problems with industrial use of symbolic regression are (1) the relatively large computational demands in comparison with other nonlinear empirical modeling techniques such as neural networks and (2) the difficulty in making the trade-off between expression accuracy and complexity. The latter issue is significant since, in general, we prefer parsimonious (simple) expressions with the expectation that they are more robust with respect to changes over time in the underlying system or extrapolation outside the range of the data used as the reference in evolving the symbolic regression.
In this chapter, we present a genetic programming variant, ParetoGP, which exploits the Pareto front to dramatically speed the symbolic regression solution evolution as well as explicitly exploit the complexity-performance trade-off. In addition to the improvement in evolution efficiency, the Pareto front perspective allows the user to choose appropriate models for further analysis or deployment. The Pareto front avoids the need to a priori specify a trade-off between competing objectives (e.g. complexity and performance) by identifying the curve (or surface or hyper-surface) which characterizes, for example, the best performance for a given expression complexity.
Keywordsgenetic programming Pareto front multi-objective optimization symbolic regression ParetoGP
Unable to display preview. Download preview PDF.
- Banzhaf, Wolfgang, Nordin, Peter, Keller, Robert E., and Francone, Frank D. (1998). Genetic Programming — An Introduction; On the Automatic Evolution of Computer Programs and its Applications. Morgan Kaufmann.Google Scholar
- Bleuler, Stefan, Brack, Martin, Thiele, Lothar, and Zitzler, Eckart (2001). Multiobjective genetic programming: Reducing bloat using SPEA2. In Proceedings of the 2001 Congress on Evolutionary Computation CEC2001, pages 536–543, COEX, World Trade Center, 159 Samseong-dong, Gangnam-gu, Seoul, Korea. IEEE Press.Google Scholar
- Castillo, Flor A., Marshall, Ken A., Green, James L., and Kordon, Arthur K. (2002). Symbolic regression in design of experiments: A case study with linearizing transformations. In Langdon, W. B., Cantú-Paz, E., Mathias, K., Roy, R., Davis, D., Poli, R., Balakrishnan, K., Honavar, V., Rudolph, G., Wegener, J., Bull, L., Potter, M. A., Schultz, A. C., Miller, J. F., Burke, E., and Jonoska, N., editors, GECCO 2002: Proceedings of the Genetic and Evolutionary Computation Conference, pages 1043–1047, New York. Morgan Kaufmann Publishers.Google Scholar
- Eckart Zitzler, Marco Laumanns and Bleuler, Stefan (2004). A tutorial on evolutionary multi-objective optimization. In Xavier Gandibleux, Marc Sevaux, Kenneth Sšrensen and T’kindt, Vincent, editors, Metaheuristics for Multiobjective Optimisation, chapter 1, pages 1–32-Springer Verlag.Google Scholar
- Jacob, Christian (2001). Illustrating Evolutionary Computation with Mathematica. Morgan Kaufmann.Google Scholar
- Kordon, Arthur, Jordaan, Elsa, Chew, Lawrence, Smits, Guido, Bruck, Torben, Haney, Keith, and Jenings, Annika (2004). Biomass inferential sensor based on ensemble of models generated by genetic programming. In in process, editor, GECCO-2004, page tbd, New York, New York.Google Scholar
- Kotanchek, Mark, Smits, Guido, and Kordon, Arthur (2003). Industrial strength genetic programming. In Riolo, Rick L. and Worzel, Bill, editors, Genetic Programming Theory and Practise, chapter 15, pages 239–256. Kluwer.Google Scholar
- Saetrom, Pal and Hetland, Magnus Lie (2003). Multiobjective evolution of temporal rules. In B. Tessem, P. Ala-Siuru, P. Doherty and Mayoh, B., editors, Proceedings of the 8th Scandinavian Conference on Artificial Intelligence. IOS Press.Google Scholar