SUMMARY

Engineering designers practice design-by-analogy (DbA) during a concept generation to retrieve knowledge or information from external sources or memory. DbA is a tool for innovation that involves retrieving analogies from a ‘source’ domain and transferring the knowledge to a ‘target’ domain. It has a history of leading to successful innovations including: Velcro inspired by the cockle burr and dog’s fur, adhesive inspired by gecko feet, and self-cleaning paint inspired by the leaves of lotus plants. Retrieval of stimuli from external sources (e.g., patent database and nature) has been significantly studied to facilitate DbA in practice. However, most studies and computational support systems have focused on a query-based approach, in which a user inputs a search term and is returned a set of algorithmically determined stimuli.

The objective of this dissertation is to approach the analogical retrieval process using a visual interaction technique so that a designer can explore for analogies, rather than be constrained by what’s retrieved by an algorithm. The study will consist of four tasks to address this overarching objective. First, algorithmic approaches will be tested, including latent semantic analysis (LSA), latent dirichlet allocation (LDA), and non-negative matrix factorization (NMF), to build a repository to enable meaningful visual exploration of US patent data for DbA inspiration. Second, visualization and exploration methods will be investigated to enable searching the repository for source analogs, including data visualization that supports a set of user interactions and re-representation of the repository based on user-defined analogical properties. Third, effects of database size on the repository and subsequent visual exploration will be characterized to assess capabilities and limitations of leveraging large-scale patent databases. Finally, a cognitive engineering design study will be conducted to evaluate the exploratory visual repository, as compared to a query-based repository, for its impact on early stage design process outcomes. The results of the proposed work will advance the field of engineering design by enhancing the way designers seek inspiration for innovation during design-by-analogy practice.