A reinforcement learning approach for reducing traffic congestion using deep Q learning.

Autor: Swapno SMMR; Department of CSE, Bangladesh University of Business and Technology, Dhaka, Bangladesh., Nobel SMN; Department of CSE, Bangladesh University of Business and Technology, Dhaka, Bangladesh., Meena P; Department of Electrical Engineering, Indian Institute of Technology, Jodhpur, Rajasthan, 342030, India., Meena VP; Department of Electrical Engineering, National Institute of Technology Jamshedpur, Jamshedpur, 831014, Jharkhand, India. vmeena1@ee.iitr.ac.in., Azar AT; College of Computer and Information Sciences, Prince Sultan University, Riyadh, 11586, Saudi Arabia. aazar@psu.edu.sa.; Automated Systems and Soft Computing Lab (ASSCL), Prince Sultan University, Riyadh, Saudi Arabia. aazar@psu.edu.sa.; Faculty of Computers and Artificial Intelligence, Benha University, Benha, 13518, Egypt. aazar@psu.edu.sa., Haider Z; College of Computer and Information Sciences, Prince Sultan University, Riyadh, 11586, Saudi Arabia.; Automated Systems and Soft Computing Lab (ASSCL), Prince Sultan University, Riyadh, Saudi Arabia., Tounsi M; College of Computer and Information Sciences, Prince Sultan University, Riyadh, 11586, Saudi Arabia. mtounsi@psu.edu.sa.; Automated Systems and Soft Computing Lab (ASSCL), Prince Sultan University, Riyadh, Saudi Arabia. mtounsi@psu.edu.sa.
Jazyk: angličtina
Zdroj: Scientific reports [Sci Rep] 2024 Dec 12; Vol. 14 (1), pp. 30452. Date of Electronic Publication: 2024 Dec 12.
DOI: 10.1038/s41598-024-75638-0
Abstrakt: Nowadays, traffic congestion is a significant issue globally. The vehicle quantity has grown dramatically, while road and transportation infrastructure capacities have yet to expand proportionally to handle the additional traffic effectively. Road congestion and traffic-related pollution have increased, which is detrimental to society and public health. This paper proposes a novel reinforcement learning (RL)-based method to reduce traffic congestion. We have developed a sophisticated Deep Q-Network (DQN) and integrated it smoothly into our system. In this study, Our implemented DQL model reduced queue lengths by 49% and increased incentives for each lane by 9%. The results emphasize the effectiveness of our method in setting strong traffic reduction standards. This study shows that RL has excellent potential to improve both transport efficiency and sustainability in metropolitan areas. Moreover, utilizing RL can significantly improve the standards for reducing traffic and easing urban traffic congestion.
Competing Interests: Declarations. Competing interests: The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.
(© 2024. The Author(s).)
Databáze: MEDLINE
Nepřihlášeným uživatelům se plný text nezobrazuje