PolyGym: Polyhedral Optimizations as an Environment for Reinforcement Learning | IEEE Conference Publication | IEEE Xplore