|
|
5. RESEARCH ISSUES
The basic premise for the need to distribute a design task is that for any reasonably sized problem, the knowledge required to solve the problem is distributed among several designers: design agents do not bring the same knowledge to bear on the design problem. In addition, there exist practical barriers that prevent a single agent from solving the problem. These barriers include geographic ones, where agents do not reside in the same physical location; and corporate ones, where agents do not reside in the same corporation. The geographic and corporate barriers result from the trend toward integrating suppliers, and even customers, into the design of large-scale products. This trend is accelerated by the availability and use of computer networks, and the popularity of the World-Wide Web. This suggests several avenues for future research.
In many cases, it is impractical, indeed impossible for many applications, to assume shared memory or data structures, such as blackboards, as a communication medium. Thus, minimal shared representations are needed to facilitate compact communication among design agents of potentially large design spaces.
The complexity of the configuration-design problems requires that effective heuristics be developed to manage the problem complexity. This chapter has proposed a CSP computational model to facilitate analysis of such heuristics, and suggested one such decomposability heuristic for constraints with certain properties (monotonicity).
Existing distributed systems focus on techniques for inter-agent negotiation and collaboration within a specific domain, or providing the infrastructure to facilitate agent interaction (Pan and Tenenbaum, 1991). More work needs to be done in this area to create generalized mechanisms for the configuration-design domain that can be applied to a wide range of problems.
6. TRENDS AND SUMMARY
With the rising popularity of the World-Wide Web and the increasing use of the Internet, the trend in design research is clearly toward facilitating interaction among teams of human designers or computer agents, or both. In this chapter, we have defined the configuration-design problem in such a way as to provide the framework for using computational design agents to solve the problem in a distributed way. We have proposed a set of specific agents for solving the problem that correspond to part catalogs and design constraints, and have identified a related set of research issues. Finally, we also show in Table 4 list of some of the more notable configuration design systems.
TABLE 4 |
A List of Some Notable Configuration Design Systems |
|
Name |
Function/task domain |
Reference |
|
MICON |
Single-board computer systems |
(Birmingham, Gupta, and Siewiorek, 1992; Gupta, Birmingham, and Siewiorek, 1993) |
ACDS |
Distributed part selection |
(Darr 1997) |
GOPS |
Optimal part selection with multi-function parts |
(Haworth, Birmingham, and Haworth, 1993) |
ParMan |
Agent-based parametric design |
(Kuokka and Livezey, 1994) |
VT |
Elevator systems |
(Marcus, Stout and McDermott, 1988) |
R1/XCON |
Computer systems |
(McDermott, 1981) |
PRIDE |
Paper-handling systems |
(Mittal, Dym, and Morjaria, 1986) |
COSSACK |
Computer systems |
(Mittal and Frayman, 1987) |
Trilogy Development Group1 |
Sales-force configuration |
(McHugh, 1996) |
|
1 Used with permission. |
|
ACKNOWLEDGMENTS
The authors are especially grateful to Professor William P. Birmingham of the Department of Electrical Engineering and Computer Science at the University of Michigan for forging our collaborative effort and commenting on our manuscript.
REFERENCES
- Birmingham, W. P., A. P. Gupta, et al. (1992). Automating the Design of Computer Systems,
- Jones and Bartlett Publishers.
- Bowen, J. and D. Bahler (1992). "Supporting multiple perspectives: a constraint-based
- approach to concurrent engineering," in Artificial Intelligence in Design 92, Kluwer Academic Publishers, Dordrecht, The Netherlands. pp. 85-96.
- Darr, T. P. (1997). A Constraint-Satisfaction Problem Computational Model forDistributed
- Part Selection, Doctoral Dissertation, Department of Electrical Engineering and Computer Science, The University of Michigan, Ann Arbor, MI.
- Deville, Y. and P. Van Hentenryck (1991). "An efficient arc consistency algorithm for a class
- of CSP problems," Twelfth International Joint Conference on Artificial Intelligence, Sydney, Australia, Morgan Kaufmann Publishers, San Mateo, CA. pp. 325-330.
- Dym, C. L. (1994). Engineering Design: A Synthesis of Views, Cambridge University Press,
- New York.
- Dym, C. L. and R. E. Levitt (1991). Knowledge-Based Systems in Engineering, McGraw-Hill,
- New York.
- Gupta, A. P., W. P. Birmingham and D. P. Siewiorek (1993). "Automating the design
- of computer systems." IEEE Transactions on Computer-Aided Design of Integrated Circuits and Systems, Vol. 12, No. 4. pp. 473-487.
- Haworth, M. S., W. P. Birmingham and D. E. Haworth (1993). "Optimal part selection."
- IEEE Transactions on Computer-Aided Design of Integrated Circuits and Systems, Vol. 12, No. 10. pp. 1611-1617.
- Kuokka, D. and B. Livezey (1994). "A collaborative parametric design agent," Twelfth
- National Conference on Artificial Intelligence, Seattle, WA. pp. 387-393.
- Mackworth, A. K. (1977). "Consistency in networks of relations," Artificial Intelligence,
- Vol. 8, No. 1. pp. 99-118.
- Marcus, S., J. Stout and J. McDermott (1988). "VT: an expert elevator designer that
- uses knowledge-based backtracking," AI Magazine, Vol. 9, No. 1. 95-112.
- McDermott, J. (1981). "R1: the formative years," AI Magazine, Vol. 2, No. 2. pp. 21-29.
- McHugh, J. (1996). "Holy Cow, No Ones Done This!," Forbes, June 3, 1996, Vol. 157, No.
- 11. pp. 122-128.
- Mittal, S., C. L. Dym, and M. Morjaria, (1986). "PRIDE: an expert system for the design
- of paper handling systems," IEEE Computer, Vol. 19, No. 7. pp. 102-114.
- Mittal, S. and F. Frayman (1987). "COSSACK: a constraints-based expert system
- for configuration tasks," Proceedings of the 2nd International Conference on Applications of AI to Engineering, Boston, MA.
- Mittal, S. and F. Frayman (1989). "Toward a generic model of configuration tasks,"
- Eleventh International Joint Conference on Artificial Intelligence (IJCAI-89), Morgan Kaufmann, Detroit, MI. pp. 1395-1401.
- Pan, J. Y. C. and J. M. Tenenbaum (1991). "An intelligent agent framework for enterprise
- integration," IEEE Transactions on Systems, Man, and Cybernetics, Vol. 21, No. 6. pp. 1391-1408.
- van Beek, P. (1992). "On the minimality and decomposability of constraint networks,"
- Tenth National Conference on Artificial Intelligence (AAAI-92). pp. 447-452.
|