RETSINA Agent Architecture



Semantic Web                Technologies

     MultiAgent Systems

     Electronic Commerce

     MAS Management

     Information Gathering
Multi-Agent Learning
MAS Interoperability


Project Grants
Site Map

The Software Agents Lab at Carnegie Mellon University's Robotics Institute envisions a world in which autonomous, intelligent software programs, known as software agents, undertake many of the operations performed by human users of the World Wide Web, as well as a multitude of other tasks. The Software Agents Lab has developed the RETSINA multi-agent system infrastructure and has applied that infrastructure and its agents to many domains, including

The RETSINA functional architecture consists of four basic agent types:

  1. Interface agents -- interact with users, receive user input, and display results.
  2. Task agents -- help users perform tasks, formulate problem-solving plans and carry out these plans by coordinating and exchanging information with other software agents.
  3. Information agents -- provide intelligent access to a heterogeneous collection of information sources.
  4. Middle agents -- help match agents that request services with agents that provide services.

Our research addresses the problem of how to facilitate communication among agents of different types. We have proposed middle or matchmaker agents to serve as liasons between agents that request services and agents that provide services. To increase interagent communication, we have also developed an agent capability description language (ACL) that allows otherwise incompatible agents to communicate.

As part of the RETSINA infrastructure of reusable agents, middle agents represent an important step in our ongoing effort to provide a foundation that will allow heterogeneous agent types and architectures to interoperate successfully. Each RETSINA agent has four reusable modules for communicating, planning, scheduling, and monitoring the execution of tasks and requests from other agents.

  • The Communication and Coordination module accepts and interprets messages and requests from other agents.
  • The Planning module takes as input a set of goals and produces a plan that satisfies the goals.

  • The Scheduling module uses the task structure created by the planning module to order the tasks.

  • The Execution module monitors this process and ensures that actions are carried out in accordance with computational and other constraints.

Below is a graphic representation of the RETSINA agent architecture:

The following pages detail our research, infrastructure, applications and publications.

Katia Sycara, our group leader and Principal Research Scientist, can be reached at nospam-katia at cs dot cmu dot edu.



Copyright 2006 - 2012 © Advanced Agent-Robotics Technology Lab - The Robotics Institute - Carnegie Mellon University

Internal Site (Restricted Access).