Scholarlink

A sustainable urban city cannot be envisioned without an efficient and reliable public transport system. From the existing infrastructure, to enhance passenger experience by reduce wait times, this research proposes a MARL based stop- skipping strategy over a heterogeneous fleet of transit vehicles. The aim is to develop intelligent control systems capable of adapting to dynamic and varied passenger demand. Modern reinforcement learning techniques are proven to be able to approximate solving complexities as in such problems, and using buses of varied capacities allows to further fulfil its potential. The results from simulation show that proposed method significantly reduce the average waiting times of the passengers.

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Stop Skipping Control over Heterogeneous Transit System Using Reinforcement Learning

  • 会议论文

  • 发表日期：

    2026年

查看全文
  • Ajeet Singh Rajput,
  • Shruti Sharma,
  • Nikita Ishwar Dhanvijay,
  • Harsh Pratap Singh,
  • Shyam Gehlot,
  • Sandeep G. Panchal

摘要

A sustainable urban city cannot be envisioned without an efficient and reliable public transport system. From the existing infrastructure, to enhance passenger experience by reduce wait times, this research proposes a MARL based stop- skipping strategy over a heterogeneous fleet of transit vehicles. The aim is to develop intelligent control systems capable of adapting to dynamic and varied passenger demand. Modern reinforcement learning techniques are proven to be able to approximate solving complexities as in such problems, and using buses of varied capacities allows to further fulfil its potential. The results from simulation show that proposed method significantly reduce the average waiting times of the passengers.