Specialization with NeuroEvolution in a collective behaviour task

ABSTRACT

In Nature, behavioral specialization is ubiquitous. Groups benefit from complementary and specialized behaviors in individuals, especially in tasks requiring collective behavior. We apply four multiagent NeuroEvolution approaches to such a task: Enforced SubPopulations [5], Parallel and Coevolutionary Enforced SubPopulations [16] and Collective NeuroEvolution [11]. Rather than just single controllers we evolve teams of simulated robots to search an unexplored area and gather certain object types for collective construction of a specific sequence. Teams are composed of agents that may evolve from initially homogeneous behavior into specialists that effectively complement each other. Results show that CONE outperforms in the collective behavior task when assisted with target behavior heuristics for lifetime learning to speed up the search. Some evolved specialists however become what we call all-rounders, taking on some more tasks to compensate for their lack in number.

REFERENCES

• 1: M. J. Baldwin. A new factor in evolution. The American Naturalist, 30(354):441--451, Jun. 1896.
• 2: C. Darwin. On The Origin of the Species by Natural Selection, or the Preservation of Favoured Races in the Struggle for Life. John Murray, London, UK, 24 Nov. 1859.
• 3: Faustino John Gomez , Risto Miikkulainen, Robust non-linear control through neuroevolution, The University of Texas at Austin, 2003
• 4: F. J. Gomez, D. Burger, and R. P. Miikkulainen. A neuroevolution method for dynamic resource allocation on a chip multiprocessor. In IJCNN'01: Proc. 2001 Int'l Joint Conf. on Neural Networks, volume 4, pages 2355--2360. IEEE Computer Society Press, Jul. 2001.
• 5: Faustino Gomez , Risto Mikkulainen, Incremental evolution of complex general behavior, Adaptive Behavior, v.5 n.3-4, p.317-342, Winter/Spring 1997 [doi>10.1177/105971239700500305]
• 6: J.-B. Lamarck. Philosophie zoologique. 1809.
• 7: Sean Luke , Claudio Cioffi-Revilla , Liviu Panait , Keith Sullivan , Gabriel Balan, MASON: A Multiagent Simulation Environment, Simulation, v.81 n.7, p.517-527, July 2005 [doi>10.1177/0037549705058073]
• 8: S. Luke, L. Panait, G. Balan, S. Paus, Z. Skolicki, J. Harrison, J. Bassett, R. Hubley, and A. Chircop. ECJ 16: A java-based evolutionary computation research system, Jan. 2008. http://cs.gmu.edu/~eclab/projects/ecj/.
• 9: J. G. Mendel. Versuche über pflanzenhybriden. Verhandlungen des naturforschenden Vereines in Brünn, 4 Abhandlungen:3--47, 1866.
• 10: David E. Moriarty , Risto Miikkulainen, Evolutionary Neural Networks for Value Ordering in Constraint SatisfactionProblems, University of Texas at Austin, Austin, TX, 1994
• 11: G. S. Nitschke , M. C. Schut , A. E. Eiben, Collective specialization in multi-rover systems, Proceedings of the 9th annual conference on Genetic and evolutionary computation, July 07-11, 2007, London, England [doi>10.1145/1276958.1277029]
• 12: D. E. Rumelhart, G. E. Hinton, and R. J. Williams. Learning internal representations by error propagation, volume 1 of Parallel Distributed Processing. MIT Press, Cambridge, USA, 1986.
• 13: A. Smith. An Inquiry into the Nature and Causes of the Wealth of Nations. Methuen and Co., Ltd. 1904, London, United Kingdom, fifth edition, 9 Mar. 1776.
• 14: P. J. Werbos. Beyond regression: New tools for prediction and analysis in the behavioral sciences. PhD thesis, Harvard University, Cambridge, MA, USA, 1974.
• 15: Ian H. Witten , Eibe Frank, Data Mining: Practical Machine Learning Tools and Techniques, Second Edition (Morgan Kaufmann Series in Data Management Systems), Morgan Kaufmann Publishers Inc., San Francisco, CA, 2005
• 16: Chern H Yong , Risto Miikkulainen, COOPERATIVE COEVOLUTION OF MULTI-AGENT SYSTEMS, University of Texas at Austin, Austin, TX, 2001

INDEX TERMS

Primary Classification:
  I. Computing Methodologies
  I.2 ARTIFICIAL INTELLIGENCE
        I.2.11 Distributed Artificial Intelligence

General Terms:
Algorithms

Keywords:
NeuroEvolution, collective behavior, specialization