Abstract
Using multiobjective genetic programming with a complexity objective to overcome tree bloat is usually very successful but can sometimes lead to undesirable collapse of the population to all single-node trees. In this paper we report a detailed examination of why and when collapse occurs. We have used different types of crossover and mutation operators (depth-fair and sub-tree), different evolutionary approaches (generational and steady-state), and different datasets (6-parity Boolean and a range of benchmark machine learning problems) to strengthen our conclusion. We conclude that mutation has a vital role in preventing population collapse by counterbalancing parsimony pressure and preserving population diversity. Also, mutation controls the size of the generated individuals which tends to dominate the time needed for fitness evaluation and therefore the whole evolutionary process. Further, the average size of the individuals in a GP population depends on the evolutionary approach employed. We also demonstrate that mutation has a wider role than merely culling single-node individuals from the population; even within a diversity-preserving algorithm such as SPEA2 mutation has a role in preserving diversity.
Similar content being viewed by others
Notes
We are indebted to an anonymous reviewer for suggesting this experiment.
References
W.B. Langdon, The evolution of size in variable length representations. In: IEEE International Conference on Evolutionary Computation, ed by P.K. Simpson (IEEE Press, Anchorage, AK, 1998), pp. 633–638
W.B Langdon, R. Poli, Fitness causes bloat: mutation. In: 1st European Workshop on Genetic Programming, ed. by W. Banzhaf, R. Poli, M. Schoenauer, T.C. Fogarty (Springer-Verlag, Paris, 1998), pp. 37–48
R. Poli, W.B. Langdon, N.F. McPhee, A Field Guide to Genetic Programming. Lulu.com (2008)
S. Silva, J. Alemida, Dynamic maximum tree depth---a simple technique for avoiding bloat in tree-based GP. In: Genetic and Evolutionary Computational Conference (GECCO 2003) (Chicago, IL, 2003), pp. 1776–1787
T. Soule, J.A. Foster, J. Dickinson, Code growth in genetic programming. In: 1st Annual Conference on Genetic Programming, ed. by J.R. Koza, D.E. Goldberg, D.B. Fogel, R.L. Riolo (MIT Pressz, Stanford University, CA, 1996), pp. 215–223
J. Stevens, R.B. Heckendorn, T. Soule, Exploiting disruption aversion to control code growth. In: Genetic And Evolutionary Computation Conference (GECCO 2005), ed. by H.-G. Beyer et al. (ACM Press, Washington DC, 2005), pp. 1605–1612
S. Luke, L. Panait, A comparison of bloat control methods for genetic programming. Evol. Comput. 14(3), 309–344 (2006)
S. Luke, L. Panait, Lexicographic parsimony pressure. In: Genetic and Evolutionary Computational Conference (GECCO 2006) (New York City, USA, 2002)
C.A.C. Coello, An updated survey of GA-based multiobjective optimization techniques. ACM Comput. Surv. 32(2), 109–143 (2000)
Y. Zhang, P.I. Rockett, Feature Extraction Using Multi-objective Genetic Programming Multi-Objective Machine Learning (Springer, Heidelberg, 2006)
Y. Zhang, P.I. Rockett. A generic Multi-dimensional Feature Extraction Method Using Multiobjective Genetic Programming. Tech. Rep. VIE 2006/002 (Department of Electronic and Electrical Engineering, University of Sheffield, Sheffield, UK, 2006)
Y. Zhang, P.I. Rockett, A Generic Optimal Feature Extraction Method Using Genetic Programming Tech. Rep. VIE 2006/001 (Department of Electronic and Electrical Engineering, University of Sheffield, Sheffield, UK, 2006)
K. Rodríguez-Vázquez, C.M. Fonseca, P.J. Fleming, Identifying the structure of non-linear dynamic systems using multiobjective genetic programming. IEEE Trans. Syst. Man Cybern. A Syst. Hum. 34(4), 531–547 (2004)
E.D de Jong, J.B Pollack, Multi-objective methods for tree size control. Genet. Program. Evol. Mach. 4(3), 211–233 (2003)
K.M.S. Badran, P.I. Rockett, The roles of diversity preservation and mutation in preventing population collapse in multiobjective genetic programming. In: Genetic and Evolutionary Computation Conference (GECCO2007), ed. by D. Theirens (ACM Press, London, UK, 2007), pp. 1551–1557
C.M. Fonseca, P.J. Fleming, Genetic algorithms for multiobjective optimization: formulation, discussion and generalization. In: 5th International Conference of Genetic Algorithms, ed. by S. Forrest (Morgan Kaufmann, San Mateo, CA, 1993), pp. 416–423
T. Ito, H. Iba, S. Sato, Depth-dependent crossover for genetic programming. In: IEEE World Congress on Computational Intelligence (IEEE Press, Anchorage, Alaska 1998), pp. 775–780
J.R. Koza, Genetic Programming: On the Programming of Computers by Means of Natural Selection (MIT Press, Cambridge, MA, 1992)
R. Kumar, P.I. Rockett, Improved sampling of the Pareto-front in multiobjective genetic optimizations by steady-state evolution: a Pareto converging genetic algorithm. Evol. Comput. 10(3), 283–314 (2002)
Y. Zhang, Multiobjective genetic programming optimal search for feature extraction. Ph.D. thesis, University of Sheffield, 2006
Y. Zhang, P.I. Rockett, Comparison of evolutionary strategies for multi-objective genetic programming. In: IEEE Systems, Man & Cybernetics Society Conference on Advances in Cybernetic Systems (AICS2006), ed. by B.P. Amavasai, N.H. Siddique, X. Cheng (Sheffield, UK, 2006)
C.L. Blake, C.J. Merz, UCI Repository of Machine Learning Databases. http://www.ics.uci.edu/mlearn/MLRepository.html, (1998).
E. Alpaydin, Combined 5 × 2 cv f test for comparing supervised classification learning algorithms. Neural Comp. 11(8), 1885–1892 (1999)
T. Dietterich, Approximate statistical tests for comparing supervised classification learning algorithms. Neural Comput. 10(7), 1895–1923 (1998)
S. Dignum, R. Poli, Generalisation of the limiting distribution of program sizes in tree-based genetic programming and analysis of its effects on bloat. In: Genetic and Evolutionary Computation Conference (GECCO2007), ed. by D. Thierens (ACM Press, London, 2007), pp. 1588–1595
K. Chellapilla, Evolving computer programs without subtree crossover. IEEE Trans. Evol. Comp. 1(3), 209–216 (1997)
E.K. Burke, S. Gustafson, G. Kendall, Diversity in genetic programming: an analysis of measures and correlation with fitness. IEEE Trans. Evol. Comput. 8(1), 47–62 (2004)
E. Zitzler, M. Laumanns, L. Thiele, SPEA2: Improving the Strength Pareto Algorithm Tech. Rep. 103 (Computer Engineering and Networks Laboratory (TIK), ETH Zurich, 2001)
K. Deb, A. Pratap, S. Agarawal, T. Meyarivan, A fast and elitist multiobjective genetic algorithm: NSGA-II. IEEE Trans. Evol. Comput. 6(2), 182–197 (2002)
R.O. Duda RO, P.E. Hart, D.G. Stork, Pattern Classification, 2nd edn. (John Wiley, 2001)
Acknowledgements
We are grateful to Dr Yang Zhang for providing the genetic programming code on which this work was based. We are also grateful to the anonymous reviewers for both suggesting and inspiring some additional experiments which have strengthened this paper.
Author information
Authors and Affiliations
Corresponding author
Rights and permissions
About this article
Cite this article
Badran, K., Rockett, P.I. The influence of mutation on population dynamics in multiobjective genetic programming. Genet Program Evolvable Mach 11, 5–33 (2010). https://doi.org/10.1007/s10710-009-9084-3
Received:
Revised:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s10710-009-9084-3