Native SAT Solver for Discrete Logic Developed

Researchers have developed Dsat, a native SAT solver for discrete logic that extends Boolean logic to handle variables with arbitrary values. Unlike traditional SAT solvers that require binarizing discrete variables into Boolean ones, Dsat operates directly on discrete variables using unit resolution and clause learning adapted for multi-valued logic. The solver is designed for applications in probabilistic reasoning, planning, and explainable AI, where discrete variables are common. Empirical comparisons show Dsat outperforms both CSP solvers on discrete CNFs and Boolean SAT solvers on binarized CNFs. The work is detailed in arXiv preprint 2605.09347.