An Opportunistic Approach to Conceptual Design
Abstract This paper addresses automation of the conceptual design stage of the overall design process, the stage that contributes the most to the cost and quality of the final product. An opportunistic constraint-based approach to conceptual design is proposed here. This approach combines some important aspects of the constraint-based transformational approach (for strongly connected systems) with aspects of the decompositional approach (for weakly connected systems) through the ability to change focus from one component to another depending on the available information and the ability to refine constraint networks hierarchically. Thus this approach enables one to address problems, such as linkage synthesis, that fall between strongly connected systems and weakly connected systems. Here design problem solving is basically viewed as an instance of the constraint satisfaction problem. Constraints are represented using hypergraphs and are manipulated through hypergraph grammars. The design strategy proposed here includes techniques to reorder the domain of a variable, to select the free choices of variables in an under-constrained system, and to select the “generator” set of constraints in an over-constrained system. A classification of constraints along with representation and reasoning guidelines is presented as a step toward implementing our approach in a way that exploits the nature of different types of constraints. An application is being developed in the linkage synthesis domain to demonstrate and validate the proposed approach.