Knowledge management of system interfaces and interactions from product development processes

Abstract

A system architecture was developed and analyzed for a basic elevator system using a limited number of system level components. A Design Structure Matrix was created which represented the complex interactions of the system components. These components were derived from a decomposition of system requirements, code and safety requirements, and evaluation of scenario operational requirements. Clustering routines using cost assignment of interactions aided in optimizing the cluster assignment of components. These cost assignments reflect cost and time associated with managing interactions inside and outside of subsystems. Management and optimization of the interfaces between the clustered components leads to an architecture that minimizes complexity and will hopefully lead to quicker and less costly product development cycles. Using this approach, near-optimal architectures can be analyzed and alternatives can be evaluated for system level impact. As was observed with this test case, highly complex or integrative systems are difficult to analyze, even with the tools utilized. These tools provided a structured approach that utilizes an objective process. This approach provides documentation and analysis of the architecture that is normally managed on the fly as product development progresses. The results of the analysis can provide a framework for an organizational structure of the product development process, provide an avenue for dialogue between design teams responsible for different subsystems, provide a process for evaluation of architecture alternatives, and identify the interactions between subsystems that must be managed carefully.