Genetic Programming and Evolvable Machines

, Volume 7, Issue 3, pp 231–252 | Cite as

Shortcomings with using edge encodings to represent graph structures

  • Gregory S. Hornby


There are various representations for encoding graph structures, such as artificial neural networks (ANNs) and circuits, each with its own strengths and weaknesses. Here we analyze edge encodings and show that they produce graphs with a node creation order connectivity bias (NCOCB). Additionally, depending on how input/output (I/O) nodes are handled, it can be difficult to generate ANNs with the correct number of I/O nodes. We compare two edge encoding languages, one which explicitly creates I/O nodes and one which connects to pre-existing I/O nodes with parameterized connection operators. Results from experiments show that these parameterized operators greatly improve the probability of creating and maintaining networks with the correct number of I/O nodes, remove the connectivity bias with I/O nodes and produce better ANNs. These results suggest that evolution with a representation which does not have the NCOCB will produce better performing ANNs. Finally we close with a discussion on which directions hold the most promise for future work in developing better representations for graph structures.


Circuits Graphs Neural networks Representations Genetic programming 



Some of this research was conducted while the author was at Sony Computer Entertainment America, Research and Development as well as Brandeis University. The soccer game and simulator was developed by Eric Larsen at SCEA R&D.


  1. 1.
    P. J. Angeline, “Morphogenic evolutionary computations: Introduction, issues and examples,” in Proc. of the Fourth Annual Conf. on Evolutionary Programming, J. McDonnell, B. Reynolds, and D. Fogel (eds.), MIT Press, 1995, pp. 387–401.Google Scholar
  2. 2.
    P. J. Angeline, G. M. Saunders, and J. B. Pollack, “An evolutionary algorithm that constructs recurrent neural networks,” IEEE Transactions on Neural Networks, vol. 5, pp. 54–65, 1994.Google Scholar
  3. 3.
    R. D. Beer and J. G. Gallagher, “Evolving dynamical neural networks for adaptive behavior,” Adaptive Behavior, vol. 1, pp. 91–122, 1992.Google Scholar
  4. 4.
    P. Bentley and S. Kumar, “Three ways to grow designs: A comparison of embryogenies of an evolutionary design problem,” in Genetic and Evolutionary Computation Conference, W. Banzhaf, J. Daida, A. E. Eiben, M. H. Garzon, V. Honavar, M. Jakiela, and R. E. Smith (eds.), Morgan Kaufmann, 1999, pp. 35–43.Google Scholar
  5. 5.
    E. J. W. Boers, H. Kuiper, B. L. M. Happel, and I. G. Sprinkhuizen-Kuyper, “Designing modular artificial neural networks,” in Proc. of Computing Science in The Netherlands, H. A. Wijshoff (ed.), SION, Stichting Mathematisch Centrum, 1993, pp. 87–96.Google Scholar
  6. 6.
    K. E. Drexler, “Biological and nanomechanical systems,” in Artificial Life, C. G. Langton (ed.), Addison Wesley, 1989, pp. 501–519.Google Scholar
  7. 7.
    F. D. Francone, M. Conrads, W. Banzhaf, and P. Nordin, “Homologous crossover in genetic programming,” in Proceedings of the Genetic and Evolutionary Computation Conference, Orlando, Florida, USA, W. Banzhaf, J. Daida, A. E. Eiben, M. H. Garzon, V. Honavar, M. Jakiela, and R. E. Smith (eds.), Morgan Kaufmann, 1999, vol. 2, pp. 1021–1026.Google Scholar
  8. 8.
    K. Fredriksson, “Genetic algorithms and generative encoding of neural networks for some benchmark classification problems,” in Proc. of the Third Nordic Workshop on Genetic Algorithms and their Applications, Finnish Artificial Intelligence Society, 1997, pp. 123–134.Google Scholar
  9. 9.
    C. M. Friedrich and C. Moraga, “An evolutionary method to find good building-blocks for architectures of artificial neural networks,” in Proc. of the Sixth Intl. Conf. on Information Processing and Management of Uncertainty in Knowledge-Based Systems, Granada, Spain, 1996, pp. 951–956.Google Scholar
  10. 10.
    F. Gruau, Neural Network Synthesis Using Cellular Encoding and the Genetic Algorithm. PhD thesis, Ecole Normale Supérieure de Lyon, 1994.Google Scholar
  11. 11.
    F. Gruau, “Automatic definition of modular neural networks,” Adaptive Behavior, vol. 3, pp. 151–183, 1995.Google Scholar
  12. 12.
    P. C. Haddow, G. Tufte, and P. van Remortel, “Shrinking the genotype: L-systems for EHW?” in Proc. of the Fourth International Conference on Evolvable Systems: From Biology to Hardware, Y. Liu, K. Tanaka, M. Iwata, T. Higuchi, and M. Yasunaga (eds.), Lecture Notes in Computer Science; Springer-Verlag, vol. 2210, 2001, pp. 128–139.Google Scholar
  13. 13.
    G. S. Hornby, Generative Representations for Evolutionary Design Automation. PhD thesis, Michtom School of Computer Science, Brandeis University, Waltham, MA, 2003.Google Scholar
  14. 14.
    G. S. Hornby, “Shortcomings with tree-structured edge encodings for neural networks,” in Proc. of the Genetic and Evolutionary Computation Conference, Vol. II, K. Deb et al. (eds.), LNCS 3103, Springer-Verlag, New York, 2004, pp. 495–506.Google Scholar
  15. 15.
    G. S. Hornby, “Measuring, enabling and comparing modularity, regularity and hierarchy in evolutionary design,” in Proc. of the Genetic and Evolutionary Computation Conference, GECCO-2005, New York, H.-G. Beyer et al. (eds.), ACM Press, 2005, pp. 1729–1736.Google Scholar
  16. 16.
    G. S. Hornby and B. Mirtich, “Diffuse versus true coevolution in a physics-based world,” in Proc. of the Genetic and Evolutionary Computation Conference, W. Banzhaf et al. (eds.), Morgan Kaufmann, 1999, pp. 1305–1312.Google Scholar
  17. 17.
    G. S. Hornby and J. B. Pollack, “Creating high-level components with a generative representation for body-brain evolution,” Artificial Life, vol. 8, pp. 223–246, 2002.Google Scholar
  18. 18.
    G. S. Hornby, S. Takamura, O. Hanagata, M. Fujita, and J. Pollack, “Evolution of controllers from a high-level simulator to a high dof robot,” in Evolvable Systems: From Biology to Hardware; Proc. of the Third Intl. Conf., J. Miller (ed.), Lecture Notes in Computer Science; Springer, vol. 1801, 2000, pp. 80–89.Google Scholar
  19. 19.
    H. Kitano, “Designing neural networks using genetic algorithms with graph generation system,” Complex Systems, vol. 4, pp. 461–476, 1990.Google Scholar
  20. 20.
    J. R. Koza, Genetic Programming: On the Programming of Computers by Means of Natural Selection. MIT Press, Cambridge, MA, 1992.Google Scholar
  21. 21.
    J. R. Koza, F. H. Bennett III, D. Andre, and M. A. Keane, “Automated WYWIWYG design of both the topology and component values of analog electrical circuits using genetic programming,” in Genetic Programming 1996: Proceedings of the First Annual Conference, Stanford University, CA, USA, J. R. Koza, D. E. Goldberg, D. B. Fogel, and R. L. Riolo (eds.), MIT Press, 1996, pp. 123–131.Google Scholar
  22. 22.
    W. B. Langdon, “Size fair and homologous tree genetic programming crossovers,” in Proceedings of the Genetic and Evolutionary Computation Conference, Orlando, Florida, USA, W. Banzhaf, J. Daida, A. E. Eiben, M. H. Garzon, V. Honavar, M. Jakiela, and R. E. Smith (eds.), Morgan Kaufmann, 1999, vol. 2, pp. 1092–1097.Google Scholar
  23. 23.
    A. Lindenmayer, “Mathematical models for cellular interaction in development,” parts I and II. Journal of Theoretical Biology, vol. 18, 1968, pp. 280–299 and 300–315.Google Scholar
  24. 24.
    J. Lohn, D. Gwaltney, G. Hornby, R. S. Zebulum, D. Keymeulen, and A. Stoica (eds.). The 2005 NASA/DoD Conference on Evolvable Hardware, Washington, DC, IEEE Press, 2005.Google Scholar
  25. 25.
    S. Luke and L. Spector, “Evolving graphs and networks with edge encoding: Preliminary report,” in Late-breaking Papers of Genetic Programming 96, Stanford Bookstore, J. Koza (ed.), 1996, pp. 117–124.Google Scholar
  26. 26.
    J. P. Mermet (ed.) IEEE Standard VHDL Language Reference Manual, IEEE Std 1076-1993, Order Code SH 16840. The Institute of Electrical and Electronics Engineers, New York, NY, 1994.Google Scholar
  27. 27.
    J. F. Miller and P. Thomson, “A developmental method for growing graphs and circuits,” in Fifth Intl. Conf. on Evolvable Systems: From Biology to Hardware, A. M. Tyrrell, P. Haddow, and J. Torrensen (eds.), Lecture Notes in Computer Science; Springer, 2003, vol. 2606, pp. 93–104.Google Scholar
  28. 28.
    S. Nolfi and D. Floreano (eds.), Evolutionary Robotics. MIT Press, Cambridge, MA, 2000.Google Scholar
  29. 29.
    S. Nolfi and D. Parisi, “growing neural networks,” Technical Report PCIA-91-15, Institute of Psychology, CNR Rome, 1992.Google Scholar
  30. 30.
    S. Nolfi and D. Parisi, “Evolving artificial neural networks that develop in time,” in European Conference on Artificial Life, F. Morán, A. Moreno, J. J. Merelo, and P. Chacón (eds.), Springer, 1995, pp. 353–367.Google Scholar
  31. 31.
    R. Poli, “Evolution of graph-like programs with parallel distributed genetic programming,” in Proc. of the Seventh Intl. Conf. on Genetic Algorithms, San Mateo, California, E. Goodman (ed.), Morgan Kaufmann, 1997, pp. 18–25.Google Scholar
  32. 32.
    R. Poli and W. B. Langdon, “Schema theory for genetic programming with one-point crossover and point mutation,” Evolutionary Computation, 1998, vol. 6, pp. 231–252.Google Scholar
  33. 33.
    A. Siddiqi and S. Lucas, “A comparison of matrix rewriting versus direct encoding for evolving neural networks,” in IEEE International Conference on Evolutionary Computation, IEEE Press, Piscataway, NJ, 1998, pp. 392–397.Google Scholar
  34. 34.
    A. Witkin and D. Baraff, “Physically based modeling: Principles and practice,” Online Siggraph '97 Course notes,∼baraff/sigcourse/index.html.
  35. 35.
    R. S. Zebulum, D. Gwaltney, G. Hornby, D. Keymeulen, J. Lohn, and A. Stoica (eds.), The 2004 NASA/DoD Conference on Evolvable Hardware, Seattle, WA, IEEE Press, 2004.Google Scholar

Copyright information

© Springer Science + Business Media, LLC 2006

Authors and Affiliations

  1. 1.University of California Santa CruzMoffett FieldUSA

Personalised recommendations