The project studies constraint-based approaches to solving combinatorial optimization problems. Such approaches represent the solutions of optimization problems as mathematical constraints and then apply an automated reasoning procedure to compute an optimal solution to the constraints. As such, constraint-based approaches can be applied to solving combinatorial optimization problems in any of the numerous domains of modern society in which they are encountered. Examples of such domains include allocating education and work opportunities, cancer treatment, matching adoptive families with children, and many problems in AI and ML. In addition to being effective, constraint-based approaches can be extended to produce formal proofs of the correctness of the solutions computed that can then be checked independently. Due to their generality, the practical optimization procedures created during the project contribute to the constantly increasing demand for trustworthy al

2024

2028