Scipediacontent: Scipediacontent moved page Draft Content 614151287 to Azevedo Rodrigues 2019a

2021-01-28T17:05:55Z

Scipediacontent moved page Draft Content 614151287 to Azevedo Rodrigues 2019a

Scipediacontent: Created page with " == Abstract == Reinforcement learning (RL) constitutes a promising solution for alleviating the problem of traffic congestion. In particular, deep RL algorithms have been sh..."

2021-01-28T17:05:52Z

Created page with " == Abstract == Reinforcement learning (RL) constitutes a promising solution for alleviating the problem of traffic congestion. In particular, deep RL algorithms have been sh..."

New page

== Abstract ==

Reinforcement learning (RL) constitutes a promising solution for alleviating the problem of traffic congestion. In particular, deep RL algorithms have been shown to produce adaptive traffic signal controllers that outperform conventional systems. However, in order to be reliable in highly dynamic urban areas, such controllers need to be robust with the respect to a series of exogenous sources of uncertainty. In this paper, we develop an open-source callback-based framework for promoting the flexible evaluation of different deep RL configurations under a traffic simulation environment. With this framework, we investigate how deep RL-based adaptive traffic controllers perform under different scenarios, namely under demand surges caused by special events, capacity reductions from incidents and sensor failures. We extract several key insights for the development of robust deep RL algorithms for traffic control and propose concrete designs to mitigate the impact of the considered exogenous uncertainties.

Comment: 8 page

== Original document ==

The different versions of the original document can be found in:

* [http://arxiv.org/abs/1904.08353 http://arxiv.org/abs/1904.08353]

* [http://arxiv.org/pdf/1904.08353 http://arxiv.org/pdf/1904.08353]

* [http://xplorestaging.ieee.org/ielx7/8907344/8916833/08917451.pdf?arnumber=8917451 http://xplorestaging.ieee.org/ielx7/8907344/8916833/08917451.pdf?arnumber=8917451],
: [http://dx.doi.org/10.1109/itsc.2019.8917451 http://dx.doi.org/10.1109/itsc.2019.8917451]

* [https://dblp.uni-trier.de/db/journals/corr/corr1904.html#abs-1904-08353 https://dblp.uni-trier.de/db/journals/corr/corr1904.html#abs-1904-08353],
: [https://arxiv.org/abs/1904.08353 https://arxiv.org/abs/1904.08353],
: [http://arxiv.org/pdf/1904.08353.pdf http://arxiv.org/pdf/1904.08353.pdf],
: [https://academic.microsoft.com/#/detail/2991611854 https://academic.microsoft.com/#/detail/2991611854]

* [https://orbit.dtu.dk/en/publications/towards-robust-deep-reinforcement-learning-for-traffic-signal-control-demand-surges-incidents-and-sensor-failures(7caba667-95b5-4836-956b-34e97e5a1fdd).html https://orbit.dtu.dk/en/publications/towards-robust-deep-reinforcement-learning-for-traffic-signal-control-demand-surges-incidents-and-sensor-failures(7caba667-95b5-4836-956b-34e97e5a1fdd).html],
: [https://orbit.dtu.dk/ws/files/194352259/1904.08353.pdf https://orbit.dtu.dk/ws/files/194352259/1904.08353.pdf]

* [ ]

← Older revision	Revision as of 17:05, 28 January 2021
(No difference)

Azevedo Rodrigues 2019a - Revision history

Scipediacontent: Scipediacontent moved page Draft Content 614151287 to Azevedo Rodrigues 2019a

Scipediacontent: Created page with " == Abstract == Reinforcement learning (RL) constitutes a promising solution for alleviating the problem of traffic congestion. In particular, deep RL algorithms have been sh..."