Towards a Knowledge-Based Model for the Computer Aided Design Process

Abstract

We propose in this paper a model for an elementary activity of the computer aided design (CAD) process, taking into account the specificity of the utilization of a computer: its capability to manipulate a relatively high volume of data and to explore alternatives quickly and systematically. The CAD activity is defined according to a given expertise domain (functionality, assembly, manufacturing, etc.) which is examined with respect to two inter-related subtasks: knowledge processing and data processing. Knowledge processing manipulates the knowledge model which represents an expert's point of view on the domain. Four things determine the knowledge model: a typology of combinable elements, a typology of possible relations between elements, an expert rule base for the creation of problem models (translation of problem data in elements and relations), and a set of criteria to evaluate the quality of solutions. Data processing manipulates the computational model which represents the computing point of view of the domain. We use the constraint satisfaction problem (CSP) model as the computational model. The relationship between subtasks is done according to the manner in which the knowledge model is translated in to the computational model. The proposed model of the CAD process allows an adaptive man-machine task sharing by allocating the user interaction to the knowledge model while the computer interacts with the computational model. An example from the assembly planning domain illustrates the model. Once the different CAD activities are determined, they can be combined to form a CAD process. The activities are coordinated by exchanging partial solutions (which can be done during knowledge processing and/or data processing), rules and evaluation criteria (during knowledge processing). Such kinds of CAD processes can also be used to define recursively the CAD activity of higher levels of abstraction.