Genetic Tradespace Exploration for Team Generation
June 2024-November 2024
This research primarily revolved around the creation and improvement of a MATLAB application that leveraged tradespace exploration for the formation of teams for a Mechanical Engineering Senior Project course. In laymen’s terms we used the creation of several generations of solutions, or team combinations assigned to different projects, and imitated the process of natural selection to propagate more successful solutions through the generations to ultimately result in better and better solutions. A better solution was defined by the level of satisfaction that all members of a team have with the members that they are partnered with as well as the project they are assigned to. After performing this calculation for a number of generations, we then displayed all the solutions in the final generation for a human user to interact with and ultimately select which solution they found the most appealing performance wise.
Containerization and Improvement of the VISION Application
December 2024-Present
The original VISION (or Visual Interaction tool for Seeking Inspiration based on Nonnegative Matrix Factorization) application seeks to do exactly what it’s name implies, provide inspiration for innovators and creators via NMF performed on a number of relevant, keyword based relationships between relevant engineering patent data. The modern VISION+ (or VISION+AI as I’ve taken to calling it) relies on artificial intelligence as it’s engine, but still wants to achieve the goal of providing inspiration to makers in a visually interactive tool.