Automated Concept Generation Using Branched Functional Models

This paper discusses a new concept generation technique that improves upon a previous automated concept generation theory and algorithm developed by Bryant, et al. at the University of Missouri – Rolla. The previous automated concept generation algorithm utilizes the design knowledge present in a repository to produce an array of partial concept solutions. While the previous algorithm is capable of handling branched functional models, it does not efficiently remove all of the infeasible partial solutions to leave only whole concepts in the final results. A matrix-based algorithm is presented in this paper that utilizes the result from the previous concept generation algorithm and solves for complete solutions of branched concepts. The presented algorithm eliminates incomplete and infeasible concepts or components from the results and generates a set of full solutions for further analysis by a designer. The details of the algorithm are described in this paper, and a peanut-sheller example is used to illustrate the effective use of the algorithm for producing branched concept variants.