A novel incremental learning scheme for reinforcement learning in dynamic environments

Zhi Wang; Chunlin Chen; Hanxiong LI; Daoyi Dong; Tzyh-Jong Tarn

doi:10.1109/WCICA.2016.7578530

A novel incremental learning scheme for reinforcement learning in dynamic environments

Zhi Wang^*
, Chunlin Chen
, Hanxiong LI
, Daoyi Dong
, Tzyh-Jong Tarn

^*Corresponding author for this work

Research output: Chapters, Conference Papers, Creative and Literary Works › RGC 32 - Refereed conference paper (with host publication) › peer-review

15 Citations (Scopus)

Abstract

In this paper, we develop a novel incremental learning scheme for reinforcement learning (RL) in dynamic environments, where the reward functions may change over time instead of being static. The proposed incremental learning scheme aims at automatically adjusting the optimal policy in order to adapt to the ever-changing environment. We evaluate the proposed scheme on a classical maze navigation problem and an intelligent warehouse system in simulated dynamic environments. Simulation results show that the proposed scheme can greatly improve the adaptability and applicability of RL in dynamic environments compared to several other direct methods.

Original language	English
Title of host publication	PROCEEDINGS OF THE 2016 12TH WORLD CONGRESS ON INTELLIGENT CONTROL AND AUTOMATION (WCICA)
Publisher	IEEE
Pages	2426-2431
Volume	2016-September
ISBN (Print)	9781467384148
DOIs	https://doi.org/10.1109/WCICA.2016.7578530
Publication status	Published - 27 Sept 2016
Event	12th World Congress on Intelligent Control and Automation, WCICA 2016 - Guilin, China Duration: 12 Jun 2016 → 15 Jun 2016

Publication series

Name
Volume	2016-September

Conference

Conference	12th World Congress on Intelligent Control and Automation, WCICA 2016
Place	China
City	Guilin
Period	12/06/16 → 15/06/16

Access Output

10.1109/WCICA.2016.7578530

Other Access Options

Check CityUHK Library

Permanent Link

Right click to copy link

Cite this

@inproceedings{68e22930fc654185be3c817daa271ef1,

title = "A novel incremental learning scheme for reinforcement learning in dynamic environments",

abstract = "In this paper, we develop a novel incremental learning scheme for reinforcement learning (RL) in dynamic environments, where the reward functions may change over time instead of being static. The proposed incremental learning scheme aims at automatically adjusting the optimal policy in order to adapt to the ever-changing environment. We evaluate the proposed scheme on a classical maze navigation problem and an intelligent warehouse system in simulated dynamic environments. Simulation results show that the proposed scheme can greatly improve the adaptability and applicability of RL in dynamic environments compared to several other direct methods.",

author = "Zhi Wang and Chunlin Chen and Hanxiong LI and Daoyi Dong and Tzyh-Jong Tarn",

year = "2016",

month = sep,

day = "27",

doi = "10.1109/WCICA.2016.7578530",

language = "English",

isbn = "9781467384148",

volume = "2016-September",

publisher = "IEEE",

pages = "2426--2431",

booktitle = "PROCEEDINGS OF THE 2016 12TH WORLD CONGRESS ON INTELLIGENT CONTROL AND AUTOMATION (WCICA)",

address = "United States",

note = "12th World Congress on Intelligent Control and Automation, WCICA 2016 ; Conference date: 12-06-2016 Through 15-06-2016",

}

Wang, Z, Chen, C, LI, H, Dong, D & Tarn, T-J 2016, A novel incremental learning scheme for reinforcement learning in dynamic environments. in PROCEEDINGS OF THE 2016 12TH WORLD CONGRESS ON INTELLIGENT CONTROL AND AUTOMATION (WCICA). vol. 2016-September, 7578530, IEEE, pp. 2426-2431, 12th World Congress on Intelligent Control and Automation, WCICA 2016, Guilin, China, 12/06/16. https://doi.org/10.1109/WCICA.2016.7578530

A novel incremental learning scheme for reinforcement learning in dynamic environments. / Wang, Zhi; Chen, Chunlin; LI, Hanxiong et al.
PROCEEDINGS OF THE 2016 12TH WORLD CONGRESS ON INTELLIGENT CONTROL AND AUTOMATION (WCICA). Vol. 2016-September IEEE, 2016. p. 2426-2431 7578530.

Research output: Chapters, Conference Papers, Creative and Literary Works › RGC 32 - Refereed conference paper (with host publication) › peer-review

TY - GEN

T1 - A novel incremental learning scheme for reinforcement learning in dynamic environments

AU - Wang, Zhi

AU - Chen, Chunlin

AU - LI, Hanxiong

AU - Dong, Daoyi

AU - Tarn, Tzyh-Jong

PY - 2016/9/27

Y1 - 2016/9/27

N2 - In this paper, we develop a novel incremental learning scheme for reinforcement learning (RL) in dynamic environments, where the reward functions may change over time instead of being static. The proposed incremental learning scheme aims at automatically adjusting the optimal policy in order to adapt to the ever-changing environment. We evaluate the proposed scheme on a classical maze navigation problem and an intelligent warehouse system in simulated dynamic environments. Simulation results show that the proposed scheme can greatly improve the adaptability and applicability of RL in dynamic environments compared to several other direct methods.

AB - In this paper, we develop a novel incremental learning scheme for reinforcement learning (RL) in dynamic environments, where the reward functions may change over time instead of being static. The proposed incremental learning scheme aims at automatically adjusting the optimal policy in order to adapt to the ever-changing environment. We evaluate the proposed scheme on a classical maze navigation problem and an intelligent warehouse system in simulated dynamic environments. Simulation results show that the proposed scheme can greatly improve the adaptability and applicability of RL in dynamic environments compared to several other direct methods.

UR - https://www.scopus.com/pages/publications/84991583825

UR - https://www.scopus.com/record/pubmetrics.uri?eid=2-s2.0-84991583825&origin=recordpage

U2 - 10.1109/WCICA.2016.7578530

DO - 10.1109/WCICA.2016.7578530

M3 - RGC 32 - Refereed conference paper (with host publication)

SN - 9781467384148

VL - 2016-September

SP - 2426

EP - 2431

BT - PROCEEDINGS OF THE 2016 12TH WORLD CONGRESS ON INTELLIGENT CONTROL AND AUTOMATION (WCICA)

PB - IEEE

T2 - 12th World Congress on Intelligent Control and Automation, WCICA 2016

Y2 - 12 June 2016 through 15 June 2016

ER -

A novel incremental learning scheme for reinforcement learning in dynamic environments

Abstract

Publication series

Conference

Access Output

Other Access Options

Permanent Link

Other Files and Links

Fingerprint

Cite this