Research will be carried out to create methods to lay out, analyze, outsource, assemble, and debug complex mechanical assemblies. To accomplish this, ways must be found to capture their fundamental structure, including the designer's strategy for locating the parts accurately with respect to each other. A concept called the `datum flow chain` will be defined and elaborated to capture this logic. Graph theory will be used to define the properties of datum flow chains. They will be explored computationally and tested on real assembly examples from the aircraft and auto industries to see how well they can capture the information designers need in order to solve typical assembly process design problems as well as to see how well they scale to large problems. The properties of different kinds of assembly features will be defined and modeled mathematically, including those that do not constrain a joint completely. A design process will be created that permits layout designers to explore the consequences for assembly and tolerances resulting from choices among different datum flow chains, feasible assembly sequences, choices of assembly mating features, and tolerances on the location and shape of these features. If successful, this research will result in increased understanding of how to model mechanical assemblies and make it easier to analyze them very early in the product design cycle. This capability could be beneficial to companies that want to improve the speed of their product development process. Improved models of assemblies could also be incorporated into engineering education, providing deeper understanding of the fundamentals of engineering systems and better preparing students to be productive in industry.
