Date of Original Version

1997

Type

Conference Proceeding

Abstract or Description

This paper presents a biologically inspired architecture for problem solving called Asynchronous Teams (A-Teams) and a Toolkit for rapid assembly and prototyping of A-Teams. A-Teams are distributed, cooperative, and scale-efficient agent-networks. We define an "agent" as anything that can act, sense, and exert some control over its actions. A-Team agents are completely autonomous, that is, each agent has exclusive control over its actions. The strengths of ATeams in problem solving arise from agent cooperation, agent distribution, and scale efficiency. Agent cooperation produces better results than can be achieved by individual agents, often leading to optimal results. A distributed architecture provides autonomy without resource constraints and control dependencies, and makes new agents relatively easy to add. Finally, scale efficiency means that the more agents that are added, the better the results in terms of solution quality and speed. The A-Teams Toolkit greatly facilitates the formation of A-Teams and provides a general software framework for distributed problem solving.

Included in

Robotics Commons

Share

COinS
 

Published In

Proceedings of the Artificial Neural Networks in Engineering (ANNIE '97), 73-78.