Automated organization design for multi-agent systems

  • Authors:
  • Mark Sims;Daniel Corkill;Victor Lesser

  • Affiliations:
  • Multi-Agent Systems Laboratory, University of Massachusetts, Amherst, USA;Multi-Agent Systems Laboratory, University of Massachusetts, Amherst, USA;Multi-Agent Systems Laboratory, University of Massachusetts, Amherst, USA

  • Venue:
  • Autonomous Agents and Multi-Agent Systems
  • Year:
  • 2008

Quantified Score

Hi-index 0.00

Visualization

Abstract

The ability to create effective multi-agent organizations is key to the development of larger, more diverse multi-agent systems. In this article we present KB-ORG: a fully automated, knowledge-based organization designer for multi-agent systems. Organization design is the process that accepts organizational goals, environmental expectations, performance requirements, role characterizations, and agent descriptions and assigns roles to each agent. These long-term roles serve as organizational-control guidelines that are used by each agent in making moment-to-moment operational control decisions. An important aspect of KB-ORG is its efficient, knowledge-informed search process for designing multi-agent organizations. KB-ORG uses both application-level and coordination-level organization design knowledge to explore the combinatorial search space of candidate organizations selectively. KB-ORG also delays making coordination-level organizational decisions until it has explored and elaborated candidate application-level agent roles. This approach significantly reduces the exploration effort required to produce effective designs as compared to modeling and evaluation-based approaches that do not incorporate design expertise. KB-ORG designs are not restricted to a single organization form such as a hierarchy, and the organization designs described here contain both hierarchical and peer-to-peer elements. We use examples from the distributed sensor network (DSN) domain to show how KB-ORG uses situational parameters as well as application-level and coordination-level knowledge to generate organization designs. We also show that KB-ORG designs effective, yet substantially different, organizations when given different organizational requirements and environmental expectations.