Abstract: This paper studies the stochastic on-time arrival (SOTA) problem in transportation networks and introduces a novel reinforcement learning-based algorithm, namely sample efficient generalized ...