Many science, engineering, and design optimization problems require balancing the trade-offs between several conflicting objectives. The objectives are often black-box functions whose evaluations are time-consuming and costly (e.g., these can be either real experiments or expensive numerical simulations). Hence, the number of experiments that can be evaluated is severely limited, and designing experiments by hand does not provide optimal performance. We propose to develop an Automated  Multi-Objective Optimal Experiment Design Platform to accelerate the discovery of optimal solutions. The platform automatically guides the design of experiments to be evaluated. Furthermore, it efficiently discovers the set of optimal solutions, called Pareto-optimal, while minimizing the number of performed evaluations. Our system provides an intuitive graphical user interface (GUI) to visualize and guide the experiments for users with little or no experience with coding, machine learning and optimization.

CSAIL Alliance members are invited to learn about this yet to be released design system from Professor Wojciech Matusik’s Computational Design and Fabrication Group. Learn more about the open-source project by visiting the summary report.