Show simple item record

dc.contributor.authorMousavi, Seyed Sajad
dc.contributor.authorschukat, Michael
dc.contributor.authorHowley, Enda
dc.date.accessioned2018-09-20T16:18:29Z
dc.date.available2018-09-20T16:18:29Z
dc.date.issued2017-08-11
dc.identifier.citationMousavi, Seyed Sajad; schukat, Michael; Howley, Enda (2017). traffic light control using deep policy-gradient and value-function based reinforcement learning . IET Intelligent Transport Systems 11 (7), 417-423
dc.identifier.issn1751-956X,1751-9578
dc.identifier.urihttp://hdl.handle.net/10379/12999
dc.description.abstractRecent advances in combining deep neural network architectures with reinforcement learning (RL) techniques have shown promising potential results in solving complex control problems with high-dimensional state and action spaces. Inspired by these successes, in this study, the authors built two kinds of RL algorithms: deep policy-gradient (PG) and value-function-based agents which can predict the best possible traffic signal for a traffic intersection. At each time step, these adaptive traffic light control agents receive a snapshot of the current state of a graphical traffic simulator and produce control signals. The PG-based agent maps its observation directly to the control signal; however, the value-function-based agent first estimates values for all legal control signals. The agent then selects the optimal control action with the highest value. Their methods show promising results in a traffic network simulated in the simulation of urban mobility traffic simulator, without suffering from instability issues during the training process.
dc.publisherInstitution of Engineering and Technology (IET)
dc.relation.ispartofIET Intelligent Transport Systems
dc.subjectgradient methods
dc.subjectlearning (artificial intelligence)
dc.subjectadaptive control
dc.subjectroad traffic control
dc.subjecttraffic engineering computing
dc.subjectcontrol engineering computing
dc.subjectdigital simulation
dc.subjecttraffic light control
dc.subjectvalue-function-based reinforcement learning
dc.subjectdeep neural network architectures
dc.subjectcomplex control problems
dc.subjecthigh-dimensional state space
dc.subjectaction spaces
dc.subjectdeep policy-gradient rl algorithm
dc.subjectvalue-function-based agent rl algorithms
dc.subjecttraffic signal
dc.subjecttraffic intersection
dc.subjectadaptive traffic light control agents
dc.subjectgraphical traffic simulator
dc.subjectcontrol signals
dc.subjectpg-based agent maps
dc.subjectoptimal control
dc.subjecturban mobility traffic simulator
dc.subjecttraining process
dc.subjectsignal control
dc.subjectfunction approximation
dc.subjectmultiagent system
dc.subjectbottlenecks
dc.subjectalgorithms
dc.subjectagent
dc.titletraffic light control using deep policy-gradient and value-function based reinforcement learning
dc.typeArticle
dc.identifier.doi10.1049/iet-its.2017.0153
dc.local.publishedsourcehttp://arxiv.org/pdf/1704.08883
nui.item.downloads0


Files in this item

This item appears in the following Collection(s)

Show simple item record