Answers are grounded in this paper and cite their sources.
This study proposes a reinforcement learning – based Variable Speed Limit (VSL) control framework for highway work zones, explicitly incorporating crash risk into decision-making. A composite reward function balancing efficiency (downstream flow) and safety (speed variation, TEE) was designed, with lane-specific VSLs optimized via a DDPG algorithm. Simulations show that, compared with no-VSL scenarios, downstream flow increased by 33.89%, while speed standard deviation and TEE decreased by 55.36% and 29.63%, respectively. Relative to an efficiency-focused strategy, the method reduced flow by 11.26% but achieved significant safety gains, including a 59.1% TEE reduction. Sensitivity analysis revealed pronounced safety – efficiency trade-offs, with γ = 0.2 and 0.3 identified as the optimal balance: a 29.63% TEE reduction under normal conditions and 28% under adverse weather, albeit with sharper flow losses. These findings underscore the robustness of the proposed framework and its potential to provide adaptive, safety-prioritized VSL strategies for highway work zones.
No takes yet. Share an insight, caveat, or question.
Wen et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: