My primary research has been in the area of Design Rationale (DR): the reasons behind making design decisions. Design Rationale differs from other forms of design documentation because it captures not only what the designer developed as a final product but also captures the reasoning behind their choices and any alternatives that were considered and rejected. DR is considered a sub-field of Artificial Intelligence in Design because it is both a Knowledge Representation problem and because AI techniques can be used in the capture and use of the rationale.

I am currently working on two related DR projects that both focus on how rationale could be used. One is the continuation of my work on the Software Engineering Using RATionale (SEURAT) system. SEURAT uses rationale to assist in software maintenance by allowing maintainers to assess the impact of changing requirements, constraints, and assumptions on an existing software system by inferencing over the previously captured rationale. SEURAT currently focuses on rationale associated with the code but is being extended to allow rationale to be associated with other development artifacts (such as design models). The other DR-based project, which I am working on with several colleagues in the Computer Science and Systems Analysis department, is Ontology-based Rationale for Collaborative Argumentation (ORCA). ORCA looks at rationale for Engineering Design and how that rationale can be used to both support design collaboration and evaluation of design decisions. The two projects, ORCA and SEURAT, are complementary since they share the same rationale representation and inference engine. More information on SEURAT can be found on the SEURAT web pages.

My Master's thesis was on Knowledge Elicitation (KE). My advisors were David C. Brown and Eva Hudlicka. This work involved building the Design Ordering Elicitation System (DOES), a Java application used to do web-based knowledge elicitation to determine the ordering of design subproblems using a combination of direct and indirect knowledge elicitation techniques. Part of this work involved creating a document classifying KE methodswhich, while not especially pretty, somehow ended up as the #1 google hit on "knowledge elicitation".

I have also worked on several SBIR projects that involved applying different AI techniques to decision support systems. Most of this work was in using genetic algorithms for battle planning, enemy action prediction, and rapid replanning. This work was performed at Charles River Analytics.

Selected Research Grants:

National Science Foundation, awarded 2009, CAREER: Rationale Capture for High Assurance Systems, $527,864

National Science Foundation, awarded 2007 (with C. Wallace (PI) and M. Siegel, Michigan Technological University; P. Anderson Miami University), CPATH-CB: The Software Communication Chautauqua, $59,794 ($34,582 Miami)

Office of the Secretary of Defense, awarded 2006 (with C. Wu (PI), Charles River Analytics, Inc.), Automated Software Analysis and Visualization, $100,000 (29,971 Miami)