Emergence of communication in competitive multi-agent systems: a pareto multi-objective approach

ABSTRACT

In this paper we investigate the emergence of communication in competitive multi-agent systems. A competitive environment is created with two teams of agents competing in an exploration task; the quickest team to explore the largest area wins. One team uses indirect communication and is controlled by an artificial neural network evolved using a Pareto multi-objective approach. The second team uses direct communication and a fixed strategy for exploration. A comparison is made between agents with and without communication. Results show that as the fitness function vary differing exploration strategies emerge. Experiments with communication produced cooperative strategies; while the experiments without communication produced effective strategies but with individuals acting independently.

REFERENCES

• 1: Hussein A. Abbass, Speeding up backpropagation using multiobjective evolutionary algorithms, Neural Computation, v.15 n.11, p.2705-2726, November 2003 [doi>10.1162/089976603322385126]
• 2: Abbass, H.A. and Deb, K. Searching under multi-evolutionary pressures. In Zitzler et al., editor, Proceedings of the Fourth Conference on Evolutionary Multi-Criterion Optimization, Spain, 2003.
• 3: Cangelosi, A. Evolution of Communication and Language Using Signals, Symbols, and Words. IEEE Transactions in Evolution Computation, 2001. 5(2): p. 93--101.
• 4: Cangelosi, A. and Parisi, D. The emergence of a 'language' in an evolving population of neural networks. Connection Science, 1998. 10: 83--97.
• 5: Elman, J.L. Finding structure in time. Cognitive Science, 14(2):179--211, 1990.
• 6: Brian P. Gerkey , Maja J. Mataric, Principled Communication for Dynamic Multi-robot Task Allocation, Experimental Robotics VII, p.353-362, December 11-13, 2000
• 7: Jones, C. and Mataric, M.J. Automatic Synthesis of Communication-Based Coordinated Multi-Robot Systems. in IEEE/RSJ International Conference on Intelligent Robots and Systems. 2004.
• 8: Simon Kirby, Natural language from artificial life, Artificial Life, v.8 n.2, p.185-215, July 2002 [doi>10.1162/106454602320184248]
• 9: MacLennan, B.J. The Emergence of Communication through Synthetic Evolution, in Advances in the Evolutionary Synthesis of Intelligent Agents, M. Patel, V. Honavar, and K. Balakrishnam, Editors. 2001, MIT Press.
• 10: Marocco, D. and Nolfi, S. Emergence of Communication in Embodied Agents Evolved for the Ability to Solve a Collective Navigation Problem, (in press).
• 11: Mataric, M.J. Using Communication to Reduce Locality in Distributed Multi-Agent Learning. in Proceedings AAAI-97. 1997. Providence, Rhode Island.
• 12: McKennoch, S., McNew, J.M. and Bushnell, L.G. A Biologically-Inspired Platform for the Evolution of Communication in Multi-Agent Systems. in Proceedings of the 2003 IEEE International Symposium on Intelligent Control. 2003. Houston, Texas.
• 13: Noble, J. Cooperation, Conflict and the Evolution of Communication. Adaptive Behavior, 1999. 7(3/4): p. 349--370.
• 14: Nolfi, S. and Marocco, D. Emergence of Communication in Embodied Agents: Co-Adapting Communicative and Non-Communicative Behaviours. In A. Cangelosi et al. (eds.), Modeling language, cognition and action: Proceedings of the 9th Neural Computation and Psychology Workshop. Singapore: World Scientific (in press).
• 15: Kenny Smith, Natural selection and cultural selection in the evolution of communication, Adaptive Behavior, v.10 n.1, p.25-44, April 2002 [doi>10.1177/10597123020101002]
• 16: Steels, L. and Kaplan, F. AIBO's first words: The social learning of language and meaning. Evolution of Communication, 2001. 4: p. 3--32.
• 17: Jason Teo , Hussein A. Abbass, Automatic generation of controllers for embodied legged organisms: a Pareto evolutionary multi-objective approach, Evolutionary Computation, v.12 n.3, p.355-394, September 2004 [doi>10.1162/1063656041774974]

INDEX TERMS

Primary Classification:
  I. Computing Methodologies
  I.2 ARTIFICIAL INTELLIGENCE
      I.2.6 Learning
          Subjects: Concept learning

Additional Classification:
  I. Computing Methodologies
  I.2 ARTIFICIAL INTELLIGENCE
      I.2.6 Learning
          Subjects: Connectionism and neural nets

General Terms:
Experimentation, Theory

Keywords:
adaptive behavior, artificial life, communication, evolutionary robotics, multi-agent systems