Using Reinforcement Learning for Per-Instance Algorithm Configuration on the TSP | IEEE Conference Publication | IEEE Xplore