关于交通安全和流动性的智能交通系统的数据驱动建模和控制解决方案
日期：2023/11/14
学术讲座：关于交通安全和流动性的智能交通系统的数据驱动建模和控制解决方案
主讲人：Shan Jiang, Lead Data Scientist, Johnson & Johnson Supply Chain, New Jersey, USA
时间：2023年11月14日（周二）上午9:00
地点: 飞书线上会议，会议链接：https://vc.feishu.cn/j/723124367
讲座摘要
Issues surrounding traffic safety and mobility have been a focal point of transportation research for decades, and they continue to affect us all. Driver behaviors, road conditions, and even weather conditions play pivotal roles in shaping the safety and mobility of our transportation systems. Moreover, the consequences of crashes and congestion extend far beyond commutes, impacting the economy, safety, and overall quality of life. Data-driven approaches are transforming our understanding of these issues, leading to effective solutions, and paving the way for better road safety, smoother traffic flow, and more efficient urban navigation. In this talk, three essential aspects of this domain will be journeyed through. First, an innovative model, ‘Safe Route Mapping’ that combines crash-based estimates with conflict risks calculated from driver-based data to score the risk associated with different roadways, will be introduced. Next, a data-driven optimization for the dynamic shortest path problem, considering time-varying travel times and traffic safety, will be explored. Lastly, the realm of large-scale adaptive traffic signal control, where a distributed multi-agent reinforcement learning with graph decomposition approach is proposed, will be ventured into. These solutions are not just theoretical—they are practical tools for improving traffic mobility, safety, and efficiency.
主讲人简介
Shan Jiang received the Ph.D. degree from the Department of Industrial and Systems Engineering, Rutgers University—New Brunswick, New Jersey, USA. His research lies in modeling, optimization and operation management of complex systems, with a focus on intelligent transportation systems, healthcare systems, and production systems. He is currently a Lead Data Scientist with Johnson & Johnson Supply Chain, New Jersey, USA, where he played a pivotal role in developing digital-twin-based automated solutions tailored for pharmaceutical manufacturing sites and supply chain management.