Euclidean Geometry by High-performance Solvers?

Abstract

Tarski showed in the 1950s that (first-order) questions in Euclidean geometry could be answered algorithmically. Algorithms for doing this have greatly improved over the decades but still have high complexity (in terms of time taken). We experiment using state-of-the-art software, specifically so-called SMT Solvers, to see how practical it is to prove classical Euclidean geometry results in this way.