Multi-Agent SystemsareSocial Systems

• Concepts from the social sciences are intrinsically relevant to MAS

  • Agent motivation via utility function maximization
  • Idea of Pareto improvements, Pareto optimality
  • Use of markets for distributed resource allocation problems (e.g., Wellman)
  • Nash equilibrium (e.g., in agent-based traffic models of Barrett and Nagel)
  • Common knowledge crucial in both game theory and MAS (Aumann, Halperin)
  • Mechanism design for automated contracting (e.g., Sandholm)

• But what happens when notions from social theory fail to appear in MAS?

Previous slide

Next slide

Back to first slide

View graphic version