为满足超多车道高速公路复杂交通管控需求，缓解道路容量不足与交通流量增长之间的矛盾，通过动态调配左侧硬路肩使用状态，实现道路空间资源优化和交通韧性提升。以广东省“一轴两网”数字化转型升级中的广深高速改扩建工程为研究对象，提出基于近端策略优化算法（PPO）的左侧硬路肩动态管控方法，并与遗传算法进行对比分析。利用SUMO仿真平台模拟不同服务水平下的事故场景，从安全和效率两个维度评估策略效果。结果表明，在一到四级服务水平下，事件条件下基于PPO算法动态开放左侧硬路肩（相较于不开放）可使总行驶时间分别降低28.8%、32.2%、31.3%、19.5%，车速标准差均值分别降低25.3%、28.1%、33.6%、28.9%，显著提升了交通系统的通行效率和行车安全性，且PPO算法效果优于传统遗传算法，为超多车道高速公路精细化管理、交通流韧性提升提供了新方法。

To address the complex traffic control requirements of super-multilane expressways and mitigate the mismatch between insufficient road capacity and growing traffic demand, this study achieves road space resource optimization and traffic resilience enhancement through dynamic regulation of the left hard shoulder. Taking the Guangzhou-Shenzhen Expressway reconstruction and expansion project under Guangdong Province′s "One Axis, Two Networks" digital transformation initiative as the research object, this study proposes a novel dynamic control method for the left hard shoulder based on the Proximal Policy Optimization (PPO) algorithm, with a comparative analysis against the Genetic Algorithm (GA). Using the SUMO simulation platform to simulate accident scenarios under different levels of service, the strategy's effectiveness is evaluated from the perspectives of safety and efficiency. The results show that under service levels one to four, dynamically opening the left hard shoulder based on the PPO algorithm under incident conditions (compared with not opening it) can reduce total travel time by 28.8%, 32.2%, 31.3%, and 19.5%, respectively, and decrease the mean speed standard deviation by 25.3%, 28.1%, 33.6%, and 28.9%, respectively. This significantly improves traffic system efficiency and driving safety, and the PPO algorithm outperforms the traditional genetic algorithm, providing a new method for refined management and traffic flow resilience enhancement on super-multilane expressways.