Reinforcement Learning and Computer Algebra

Performing Buchberger's algorithm, the standard algorithm for computing Groebner bases, involves making many important choices. In particular, the strategy for selecting the next s-pair to process can make dramatic differences in the time and space needed for the computation. While we have several heuristics for pair selection, reinforcement learning provides tools for exploring and developing new strategies for both Buchberger's algorithm and other algorithms in computer algebra.