Hi-CoLA: High Confidence Lower Bound Approximation Based Reinforcement Learning for Flex-Route Transit Operation Control
Hi-CoLA: High Confidence Lower Bound Approximation Based Reinforcement Learning for Flex-Route Transit Operation Control
This research introduces Hi-CoLA, a reinforcement learning framework designed for safe policy improvement in real-time flex-route transit (FRT) operation control. By combining importance sampling, concentration inequalities, and a differentiable surrogate network, the method efficiently estimates high-confidence performance lower bounds to optimize robustness. The study shows that Hi-CoLA outperforms state-of-the-art offline and safe RL methods in both average and tail-risk metrics within a case study on an actual Chicago route implemented in the SUMO simulator, offering a proof of concept for safety-critical decision-making scenarios.
The paper proposes Hi-CoLA, a reinforcement learning framework that uses importance sampling and concentration inequalities to estimate high-confidence performance lower bounds for flex-route transit control, demonstrating improved average and tail-risk performance in SUMO simulations of a Chicago route.
Source: arXiv Robotics — research abstracts · Read original article ↗
Loading article text…
What the source reports
Publisher-reported claims, with original evidence. These results have not been independently verified by RoboSignal.
- Environment
- SimulationOpen source S4
- paper: Reported pending Open source S2
- paper: not_reported
Source excerpts and review record
Automatically extracted; no manual editorial approval recorded.
ement. On the FRT operation control problem, Hi-CoLA outperforms state-of-the-art offline and safe RL methods in both average and tail-risk performance. A case study on an actual route in Chicago, implemented in the SUMO simulator, serves as a proof of concept. Hi-CoLA enhances real-time FRT control with guaranteed performance and is broadly applicable to other safety critical decision-making scenarios.
Open source S4
arXiv:2611.01071v1 Announce Type: new Abstract: This paper studies safe policy improvement for real-time operation control of flex-route transit (FRT) systems, where robustness is critical and any performance decrease is undesirable. While reinforcement learning (RL) has shown strong empirical success, its adoption remains limited in safety and cost critical applications, such as intelligent transportation, since lea
Open source S2
Source：arXiv Robotics — research abstracts · arxiv.org