Optimizing Transportation Dynamics at a City-Scale Using a Reinforcement Learning Framework

Khaidem, Luckyson; Luca, Massimiliano; Yang, Fan; Anand, Ankit; Lepri, Bruno; Dong, Wen

doi:10.1109/access.2020.3024979

Cited by 14 publications

(4 citation statements)

References 34 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Simulations are extensively used in a variety of fields such as urban planning [25], transportation [26]- [28], robotics [29], epidemiology [30], gaming [31], and others. Experiments in the real world are often costly, dangerous, and infeasible.…”

Section: Background a Simulations In Sumomentioning

confidence: 99%

Autonomous and Human-Driven Vehicles Interacting in a Roundabout: A Quantitative and Qualitative Evaluation

Ferrarotti,

Luca,

Santin

et al. 2024

IEEE Access

View full text Add to dashboard Cite

Optimizing traffic dynamics in an evolving transportation landscape is crucial, particularly in scenarios where autonomous vehicles (AVs) with varying levels of autonomy coexist with human-driven cars. While optimizing Reinforcement Learning (RL) policies for such scenarios is becoming more and more common, little has been said about realistic evaluations of such trained policies. This paper presents an evaluation of the effects of AVs penetration among human drivers in a roundabout scenario, considering both quantitative and qualitative aspects. In particular, we learn a policy to minimize traffic jams (i.e., minimize the time to cross the scenario) and to minimize pollution in a roundabout in Milan, Italy. Through empirical analysis, we demonstrate that the presence of AVs can reduce time and pollution levels. Furthermore, we qualitatively evaluate the learned policy using a cutting-edge cockpit to assess its performance in nearreal-world conditions. To gauge the practicality and acceptability of the policy, we conduct evaluations with human participants using the simulator, focusing on a range of metrics like traffic smoothness and safety perception. In general, our findings show that human-driven vehicles benefit from optimizing AVs dynamics. Also, participants in the study highlight that the scenario with 80% AVs is perceived as safer than the scenario with 20%. The same result is obtained for traffic smoothness perception.

show abstract

Section: Background a Simulations In Sumomentioning

confidence: 99%

Autonomous and Human-Driven Vehicles Interacting in a Roundabout: A Quantitative and Qualitative Evaluation

Ferrarotti,

Luca,

Santin

et al. 2024

IEEE Access

View full text Add to dashboard Cite

show abstract

“…Their experiments also show its strength in computing time, system efficiency, and robustness. To further enhance the generality, (Khaidem et al 2020) coped with the difficulty of gathering insights for an entire city through a Partially Observable Discrete Event Decision Process (PODEDP).…”

Section: Reinforcement Learning In Traffic and Transit Operationsmentioning

confidence: 99%

A Reinforcement Method for Passenger Flow Control Based on Asynchronous Advantage Actor-Critic Neural Network

Wang

Jin

Luo

et al. 2023

Preprint

View full text Add to dashboard Cite

Effective passenger flow management is critical for improving service quality and alleviating congestion in metro networks. However, the dynamic nature of travel demand and the complex structure of metro networks present significant challenges in building and solving control models. Additionally, the high computational costs of existing methods limit their practical applications. To address these challenges, this study proposes a new reinforcement learning (RL) based method for passenger flow control. The method has three components: the network state characterization, the control model, and the reinforcement learning model. Then, the study outlines the “action”, “state”, and “reward” concepts in RL based on the definition of decision variables, constraints, and objective functions in the constructed passenger flow control programming model. An iterative interaction mechanism is introduced to synchronize the control schemes generated by the reinforcement learning unit and the network states. Furthermore, effectively utilizing computational resources, the Asynchronous Advantage Actor-Critic Neural Network (A3C-NN) is trained to optimize the complex programming model. Finally, the proposed approach is validated through a case study using data from Chengdu Urban Rail Transit (URT), demonstrating its effectiveness in achieving various objectives, such as minimizing passenger waiting time, maximizing passenger turnover, and maximizing passenger numbers.

show abstract

“…𝑖,𝑗 T 𝑖 𝑗 + 𝑖 𝑗 𝑇 𝑖 𝑗 (14) where T𝑖 𝑗 is the flow from region 𝑖 to region 𝑗 predicted by the model and 𝑇 𝑖 𝑗 is the actual flow from region 𝑖 to region 𝑗. CPC ranges between 0 and 1: if two adjacency matrices do not have any flows in common, CPC value is 0. CPC is 1 if the sets of flows are identical.…”

Section: Evaluation Metricsmentioning

confidence: 99%

Enhancing crowd flow prediction in various spatial and temporal granularities

Cardia¹,

Luca²,

Pappalardo³

2022

Preprint

Self Cite

View full text Add to dashboard Cite

Thanks to the diffusion of the Internet of Things, nowadays it is possible to sense human mobility almost in real time using unconventional methods (e.g., number of bikes in a bike station). Due to the diffusion of such technologies, the last years have witnessed a significant growth of human mobility studies, motivated by their importance in a wide range of applications, from traffic management to public security and computational epidemiology. A mobility task that is becoming prominent is crowd flow prediction, i.e., forecasting aggregated incoming and outgoing flows in the locations of a geographic region. Although several deep learning approaches have been proposed to solve this problem, their usage is limited to specific types of spatial tessellations and cannot provide sufficient explanations of their predictions. We propose CrowdNet, a solution to crowd flow prediction based on graph convolutional networks. Compared with state-of-the-art solutions, CrowdNet can be used with regions of irregular shapes and provide meaningful explanations of the predicted crowd flows. We conduct experiments on public data varying the spatio-temporal granularity of crowd flows to show the superiority of our model with respect to existing methods, and we investigate CrowdNet's reliability to missing or noisy input data. Our model is a step forward in the design of reliable deep learning models to predict and explain human displacements in urban environments.

show abstract

Optimizing Transportation Dynamics at a City-Scale Using a Reinforcement Learning Framework

Cited by 14 publications

References 34 publications

Autonomous and Human-Driven Vehicles Interacting in a Roundabout: A Quantitative and Qualitative Evaluation

Autonomous and Human-Driven Vehicles Interacting in a Roundabout: A Quantitative and Qualitative Evaluation

A Reinforcement Method for Passenger Flow Control Based on Asynchronous Advantage Actor-Critic Neural Network

Enhancing crowd flow prediction in various spatial and temporal granularities

Contact Info

Product

Resources

About