Scholarlink

<p>Vehicular traffic intensity impacts public health and environmental sustainability due to congestion and the emission of harmful gases. As vehicular traffic increases on Vehicular Road Networks (VRNs), real time monitoring and trajectory optimization become essential. Vehicular trajectory path optimization has been addressed by employing different algorithms such as Q-Learning, Ant Colony Optimization (ACO), and the Firefly Optimization Algorithm (FO). However, using each of these algorithms individually presents limitations in vehicular environments. Q-learning has slow convergence, ACO are susceptible to becoming trapped in local optima, and FO are sensitive to parameter tuning and often inefficient in small space exploration. Therefore, to address these challenges, a novel hybrid framework, AIQAFO, is proposed, which provides adaptive global decision-making, cooperative local density estimation, and enhanced traffic distribution capabilities simultaneously. The proposed framework improves Quality of Service (QoS) parameters (trip duration, distance traveled, fuel consumption, and <InlineEquation ID="IEq1"> <EquationSource Format="TEX">\(CO_2\)</EquationSource> </InlineEquation> emissions) compared to various state of the art algorithms. The results were benchmarked against the Shortest Time Routing (STR) approach as a baseline, the Pheromone Based Routing (PheR), Repelling Pheromone Based Rerouting (RepelPheRR), and Dynamic Traffic Assignment (DTA) approaches. By effectively reducing traffic congestion, the AIQAFO method achieves a 51.29 % reduction in journey time, a 7% decrease in total distance traveled, a 39% reduction in fuel consumption, and a 26% decrease in carbon dioxide emissions compared to the baseline STR. QoS improvements highlight the framework’s effectiveness in promoting both environmental sustainability and enhanced vehicular mobility.</p>

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

Q-Learning Based Nature Inspired Algorithm for Traffic Optimization in Vehicular Networks

  • 期刊论文

  • 发表日期：

    2025年10月20日

查看全文
  • Piyush Chauhan,
  • Nishant Sharma,
  • Alok Kumar

摘要

Vehicular traffic intensity impacts public health and environmental sustainability due to congestion and the emission of harmful gases. As vehicular traffic increases on Vehicular Road Networks (VRNs), real time monitoring and trajectory optimization become essential. Vehicular trajectory path optimization has been addressed by employing different algorithms such as Q-Learning, Ant Colony Optimization (ACO), and the Firefly Optimization Algorithm (FO). However, using each of these algorithms individually presents limitations in vehicular environments. Q-learning has slow convergence, ACO are susceptible to becoming trapped in local optima, and FO are sensitive to parameter tuning and often inefficient in small space exploration. Therefore, to address these challenges, a novel hybrid framework, AIQAFO, is proposed, which provides adaptive global decision-making, cooperative local density estimation, and enhanced traffic distribution capabilities simultaneously. The proposed framework improves Quality of Service (QoS) parameters (trip duration, distance traveled, fuel consumption, and \(CO_2\) emissions) compared to various state of the art algorithms. The results were benchmarked against the Shortest Time Routing (STR) approach as a baseline, the Pheromone Based Routing (PheR), Repelling Pheromone Based Rerouting (RepelPheRR), and Dynamic Traffic Assignment (DTA) approaches. By effectively reducing traffic congestion, the AIQAFO method achieves a 51.29 % reduction in journey time, a 7% decrease in total distance traveled, a 39% reduction in fuel consumption, and a 26% decrease in carbon dioxide emissions compared to the baseline STR. QoS improvements highlight the framework’s effectiveness in promoting both environmental sustainability and enhanced vehicular mobility.