AN OBJECT-ORIENTED REPRESENTATION FOR PRODUCT AND DESIGN PROCESSES

We report on the development of a knowledge representation model, whic h is based on the SHARED object model reported in Shared Workspaces fo r Computer-Aided Collaborative Engineering (Wong, A. and Sriram, D., T echnical Report, IESL 93-06, Intelligent Engineering Systems Laborator y, Department of Civil Engineering, MIT, March, 1993) and Research in Engineering Design (Wong, A. and Sriram, D., SHARED: An Information Mo del for Cooperative Product Development, 1993, Fall, 21-39). Our curre nt model is implemented as a layered scheme, that incorporates both an evolving artifact and its associated design process. To represent art ifacts as they evolve, we define objects recursively without a pre-def ined granularity on this recursive decomposition. This eliminates the need for translations between levels of abstraction in the design proc ess. The SHARED model extends traditional OOP in three ways: (1) by al lowing explicit relationship classes with inheritance hierarchies; (2) by permitting constraints to be associated with objects and relations hips; and (3) by comparing 'similar' objects at three different levels (form, function and behavior). Five primitive objects define the desi gn process: goal, plan, specification, decision and context. Goal obje cts achieve function, introduce constraints, introduce new artifacts o r modify existing ones, and create subgoals. Plan objects order goals and link a product hierarchy to a process hierarchy. Specification obj ects define user inputs as constraints. Decision objects relate goals to user decisions and context objects describe the design context. Ope rators that are applied to design objects collectively form a represen tation of the design process for a given context. The representation i s robust enough to effectively model four design paradigms [described in Journal of CAD (Gorti, S. and Sriram, R. D., Symbol to Form Mapping : a Framework for Conceptual Design, 1996, 28(11), 853-870)]: top-down decomposition, stepwise refinement, bottom-up composition and constra int propagation. To demonstrate this, we represent the designs of two TV remote controllers in the SHARED architecture. The example reveals that certain aspects of artifact knowledge are essentially context-ind ependent and that this representation can be a foundation for robust k nowledge-based systems in design. (C) 1998 Published by Elsevier Scien ce Ltd. All rights reserved.