On Evolving Social Systems: Communication, Speciation and Symbiogenesis

  • Authors:
  • Larry Bull

  • Affiliations:
  • Intelligent Computer Systems Centre, University of the West of England, Bristol BS16 1QY, UK. larry@ics.uwe.ac.uk

  • Venue:
  • Computational & Mathematical Organization Theory
  • Year:
  • 1999

Quantified Score

Hi-index 0.00

Visualization

Abstract

In this paper we introduce three enhancements forevolutionary computing techniques in social environments. We describethe use of the genetic algorithm to evolve communicating rule-basedsystems, where each rule-based system represents an agent in asocial/multi-agent environment. It is shown that the evolution ofmultiple cooperating agents can give improved performance over theevolution of an equivalent single agent, i.e. non-social, system. Weexamine the performance of two social system configurations asapproaches to the control of gait in a wall climbing quadrupedalrobot, where each leg of the quadruped is controlled by acommunicating agent. We then introduce two social-leveloperators—speciation and symbiogenesis—which aim to reduce theamount of knowledge required a priori by automatically manipulatingthe system‘s social structure and describe their use in conjunctionwith the communicating rule-based systems. The reasons forimplementing these kinds of operators are discussed and we thenexamine their performance in developing the controller of thewall-climbing quadruped. We find that the use of such operators cangive improved performance over static population/agent configurations.