Bias and scalability in evolutionary development

ABSTRACT

The introduction of a genotype-phenotype map modelled on biological development can potentially improve the scalability of evolutionary algorithms. Previous work by Gordon and Bentley demonstrated that such a model can be used to evolve patterns that map to useful but small phenotypes. This paper uses the same model to generate much larger patterns covering arrays of up to 64x64 cells. The results show that the model's performance is generally comparable to similar development-based systems [12, 14], and with some measures outperforms them. Additionally the inherent biases of the model are explored, such as the need to use symmetry-breaking initial conditions which some other models do not require. This exploration yields a set of guidelines that suggest what kinds of problem the model is suited to exploring.

REFERENCES

• 1: Bongard, J.C. Evolving Modular Genetic Regulatory Networks. 2002 Congress on Evolutionary Computation. 2002. Honolulu, HI, U.S.A.: IEEE Press. pp. 1872--1877.
• 2: Cotton, F.A., Chemical Applications of Group Theory. 2nd ed. 1990, New York: John Wiley & Sons Inc. 478.
• 3: Dawkins, R. The evolution of evolvability. Proc. of Artificial Life. 1989. Santa Fe, U.S.A.: Addison-Wesley. pp. 201--220.
• 4: Peter Eggenberger, Creation of Neural Networks Based on Developmental and Evolutionary Principles, Proceedings of the 7th International Conference on Artificial Neural Networks, p.337-342, October 08-10, 1997
• 5: Gordon, T.G.W. Exploring Models of Development for Evolutionary Circuit Design. Congress on Evolutionary Computation. 2003. Canberra, Australia: IEEE Press, Piscataway, NJ, USA. pp. 2050--2057.
• 6: Timothy G. W. Gordon , Peter J. Bentley, Towards Development in Evolvable Hardware, Proceedings of the 2002 NASA/DoD Conference on Evolvable Hardware (EH'02), p.241, July 15-18, 2002
• 7: Gruau, F., Cellular Encoding of Genetic Neural Network. 1992, Laboratoire de lInformatique pour le Parallelisme, Ecole Normale Superieure de Lyon: Lyon.
• 8: Pauline C. Haddow , Gunnar Tufte , Piet van Remortel, Shrinking the Genotype: L-systems for EHW?, Proceedings of the 4th International Conference on Evolvable Systems: From Biology to Hardware, p.128-139, October 03-05, 2001
• 9: Hornby, G.S. and J.B. Pollack. The advantages of generative grammatical encodings for physical design. Proc. Congress on Evolutionary Computation. 2001. Seoul, South Korea.
• 10: Jakobi, N., Harnessing Morphogenesis, in On Growth, Form and Computers, S. Kumar and P.J. Bentley, Editors. 2003, Elsevier: London.
• 11: John R. Koza , Martin A. Keane , Matthew J. Streeter, The Importance of Reuse and Development in Evolvable Hardware, Proceedings of the 2003 NASA/DoD Conference on Evolvable Hardware, p.33, July 09-11, 2003
• 12: Miller, J.F. Evolving a self-repairing, self-regulating, French flag organism. Genetic and Evolutionary Computation Conf.. 2004. Seattle, Washington, USA.
• 13: Julian F. Miller , Peter Thomson, Cartesian Genetic Programming, Proceedings of the European Conference on Genetic Programming, p.121-132, April 15-16, 2000
• 14: Roggen, D. and D. Federici. Multi-cellular Development: Is There Scalability and Robustness to Gain? 8th Int. Conf. on Parallel Problem Solving in Nature. 2004. Birmingham, U.K. pp. 391--400.
• 15: A new kind of science, Wolfram Media Inc., Champaign, Ilinois, US, 2002
• 16: Wolpert, L., et al., Principles of Development. 2nd Edition ed. 2002, Oxford, U.K.: Oxford University Press.
• 17: Yao, X. and T. Higuchi, Promises and challenges of evolvable hardware. IEEE Transactions on Systems, Man and Cybernetics, Part C Applications and Reviews, 1999. 29(1): pp. 87--97.

INDEX TERMS

Primary Classification:
  I. Computing Methodologies
  I.2 ARTIFICIAL INTELLIGENCE
      I.2.2 Automatic Programming
          Subjects: Program synthesis

General Terms:
Algorithms, Experimentation, Performance

Keywords:
development, pattern formation, scalability