Innovation Mining: A framework for identifying components worth innovating in a system
Srikanth Peyyeti · RIT Scholar Works (Rochester Institute of Technology) · 2016
With advancement in technology and increased breadth of requirements of customers, organizations have to deliver innovative products in a quick and efficient way to remain competitive. While there are many theories and approaches to innovation in products, few of them are quantitative in nature. Furthermore, most product development starts from an existing system. Thus, a starting assumption in this work is that a system architecture and design already exists. The main goals of this research are to develop a framework that identifies innovation opportunities within an existing product and to develop a quantitative approach for the identification of those opportunities. In this work, three scenarios are defined that signal the need for innovation. These scenarios are (1) the detection of a leveling off of the benefit-to-cost ratio of the system over time (i.e. the S-Curve), (2) trade-offs between technical parameters within the product and (3) a new need for the product has been identified. Using these indicators as a basis, this thesis focuses on identifying the physical components that are most likely to lead to innovation. Once the relevant engineering metrics associated with these scenarios are identified, they are deployed through the product system links. The framework proposed in this thesis leverages the relationships between requirements and product structure, which are represented as matrices and then singular value decomposition (SVD) and clustering analysis techniques are used to identify patterns within the requirements and the components to focus the innovation efforts of the product developers. This framework is applied to a case study to assess the initial feasibility of the approach.