Benchmarks for reinforcement learning in mixed-autonomy traffic

Eugene Vinitsky, Aboudy Kreidieh, Luc Le Flem, Nishant Kheterpal, Kathy Jang, Cathy Wu, Fangyu Wu, Richard Liaw, Eric Liang, Alexandre M. Bayen
Proceedings of The 2nd Conference on Robot Learning, PMLR 87:399-409, 2018.

Abstract

We release new benchmarks in the use of deep reinforcement learning (RL) to create controllers for mixed-autonomy traffic, where connected and autonomous vehicles (CAVs) interact with human drivers and infrastructure. Benchmarks, such as Mujoco or the Arcade Learning Environment, have spurred new research by enabling researchers to effectively compare their results so that they can focus on algorithmic improvements and control techniques rather than system design. To promote similar advances in traffic control via RL, we propose four benchmarks, based on three new traffic scenarios, illustrating distinct reinforcement learning problems with applications to mixed-autonomy traffic. We provide an introduction to each control problem, an overview of their MDP structures, and preliminary performance results from commonly used RL algorithms. For the purpose of reproducibility, the benchmarks, reference implementations, and tutorials are available at https://github.com/flow-project/flow.

Cite this Paper

BibTeX
@InProceedings{pmlr-v87-vinitsky18a, title = {Benchmarks for reinforcement learning in mixed-autonomy traffic}, author = {Vinitsky, Eugene and Kreidieh, Aboudy and Flem, Luc Le and Kheterpal, Nishant and Jang, Kathy and Wu, Cathy and Wu, Fangyu and Liaw, Richard and Liang, Eric and Bayen, Alexandre M.}, booktitle = {Proceedings of The 2nd Conference on Robot Learning}, pages = {399--409}, year = {2018}, editor = {Billard, Aude and Dragan, Anca and Peters, Jan and Morimoto, Jun}, volume = {87}, series = {Proceedings of Machine Learning Research}, month = {29--31 Oct}, publisher = {PMLR}, pdf = {http://proceedings.mlr.press/v87/vinitsky18a/vinitsky18a.pdf}, url = {https://proceedings.mlr.press/v87/vinitsky18a.html}, abstract = {We release new benchmarks in the use of deep reinforcement learning (RL) to create controllers for mixed-autonomy traffic, where connected and autonomous vehicles (CAVs) interact with human drivers and infrastructure. Benchmarks, such as Mujoco or the Arcade Learning Environment, have spurred new research by enabling researchers to effectively compare their results so that they can focus on algorithmic improvements and control techniques rather than system design. To promote similar advances in traffic control via RL, we propose four benchmarks, based on three new traffic scenarios, illustrating distinct reinforcement learning problems with applications to mixed-autonomy traffic. We provide an introduction to each control problem, an overview of their MDP structures, and preliminary performance results from commonly used RL algorithms. For the purpose of reproducibility, the benchmarks, reference implementations, and tutorials are available at https://github.com/flow-project/flow.} }
Endnote
%0 Conference Paper %T Benchmarks for reinforcement learning in mixed-autonomy traffic %A Eugene Vinitsky %A Aboudy Kreidieh %A Luc Le Flem %A Nishant Kheterpal %A Kathy Jang %A Cathy Wu %A Fangyu Wu %A Richard Liaw %A Eric Liang %A Alexandre M. Bayen %B Proceedings of The 2nd Conference on Robot Learning %C Proceedings of Machine Learning Research %D 2018 %E Aude Billard %E Anca Dragan %E Jan Peters %E Jun Morimoto %F pmlr-v87-vinitsky18a %I PMLR %P 399--409 %U https://proceedings.mlr.press/v87/vinitsky18a.html %V 87 %X We release new benchmarks in the use of deep reinforcement learning (RL) to create controllers for mixed-autonomy traffic, where connected and autonomous vehicles (CAVs) interact with human drivers and infrastructure. Benchmarks, such as Mujoco or the Arcade Learning Environment, have spurred new research by enabling researchers to effectively compare their results so that they can focus on algorithmic improvements and control techniques rather than system design. To promote similar advances in traffic control via RL, we propose four benchmarks, based on three new traffic scenarios, illustrating distinct reinforcement learning problems with applications to mixed-autonomy traffic. We provide an introduction to each control problem, an overview of their MDP structures, and preliminary performance results from commonly used RL algorithms. For the purpose of reproducibility, the benchmarks, reference implementations, and tutorials are available at https://github.com/flow-project/flow.
APA
Vinitsky, E., Kreidieh, A., Flem, L.L., Kheterpal, N., Jang, K., Wu, C., Wu, F., Liaw, R., Liang, E. & Bayen, A.M.. (2018). Benchmarks for reinforcement learning in mixed-autonomy traffic. Proceedings of The 2nd Conference on Robot Learning, in Proceedings of Machine Learning Research 87:399-409 Available from https://proceedings.mlr.press/v87/vinitsky18a.html.

Related Material