Ai-Mei Chang

Professor of Systems Management
National Defense University
Information Resources Management College
Ft. McNair
Washington, DC 20319
e-mail: chang@ndu.edu
phone: 202-685-4889

Award Number: IRI9409924

Cooperation Support Systems for Group Problem Solving

Project Summary

Many problems that are encountered in an organizational setting are large and complex enough that they require a team/group effort for efficient and effective resolution. These problems are usually decomposed into subproblems which allow team members, eac h of whom is responsible for a distinct subproblem depending on their area of expertise, to work on subproblems and resolve the overall problem through coordination with each other. In some cases, the problems are well-structured that a decomposition of t he overall problem will result in relatively independent subproblems, which may facilitate concurrent resolution of subproblems. This is followed by information exchange and cooperation between team members to resolve the overall problem. In some cases, p roblems may be ill-structured, ill-defined, and/or subproblems highly dependent that team members may have to collectively deliberate on the various aspects of the problem to resolve it. Most other problems are usually in between these extreme cases. Exa mples of such problems range from engineering and software design, project planning to project selection, product design and marketing, make or buy decisions, audit going concern decisions, and many other organizational decisions that involve an interface of different organizational functions. This project addresses the issue of designing information sharing and cooperation mechanisms to support the overall problem resolution process using AI techniques and Argumentation Theory.

The design of the cooperation support system is based on two concepts: a synthetic process of subproblem resolution at the individual agent level and an analytical process of overall problem resolution at the group level. The synthetic process involves a consistent integration of premises, assumptions, and rules to arrive at propositions regarding the subproblem, which is accomplished using AI techniques such as Assumption-Based Truth Maintenance Systems. The analytical process at the group level depends on the outputs of the synthetic processes and is accomplished using techniques based on Argumentation Theory. The overall problem resolution process is shown to be an iterative process of synthesis and analysis.

A prototype system (using Logic-Based Truth Maintenance System/ Assumption Based Truth maintenance System) has been developed based on the above design for solving group problems in a specific domain, namely auditing. This prototype serves as an illustra tion of the viability and effectiveness of our design in

  1. better understanding the group problem solving processes in domain-specific activities and in
  2. building effective and efficient cooperation systems in supporting such activities by incorporating the coordination mechanisms that are specific to the domain.
The prototype forms the basis for extensive ongoing experimentation on the efficiency and efficacy of such systems using field studies in auditing and other organizations. The argumentation support system has also been designed. The application area mod eled for the prototype system is purchase negotiations. The system allows multiple users to access the same knowledge base and update it. It also includes expert knowledge of the purchase domain to guide the argumentation process

The project provides substantive insights into the nature of domain-specific coordination processes and mechanisms and how to effectively incorporate them into the software environment while designing cooperation support systems. This can facilitate orga nizational control of collaborative processes enabling efficient and effective cooperation.

References


Return to ITO Workshop Abstracts

Return to ITO Workshop Home Page