A Multi-Step Reinforcement Learning Algorithm

Zhi Cong Zhang; Kai Shun Hu; Hui Yu Huang; Shuai Li; Shao Yong Zhao

doi:10.4028/www.scientific.net/AMM.44-47.3611

Paper Titles

A Kind of Automatic Verification Apparatus for Dial Indicator
p.3589

Research on Path Planning of Mine Rescue Robots Based on Fuzzy Control
p.3593

Research on Flame Simulation Based on Improved Particle System and the Texture Mapping
p.3601

Development of a Unified Resource Base for E-Learning
p.3606

A Multi-Step Reinforcement Learning Algorithm
p.3611

The Effect of Zirconium Incorporation on the Brønsted Acidity of Zeolite: A DFT Study
p.3616

Study on the Dynamic Characteristics Analysis and Failure Prognosis of a PWA
p.3620

Data Acquisition and Processing System for Vehicle's Braking Performance Based on Visual C++
p.3627

Study of Pavement Identification Approach Based on Wavelet Analysis
p.3632

HomeApplied Mechanics and MaterialsApplied Mechanics and Materials Vols. 44-47A Multi-Step Reinforcement Learning Algorithm

A Multi-Step Reinforcement Learning Algorithm

Abstract:

Reinforcement learning (RL) is a state or action value based machine learning method which approximately solves large-scale Markov Decision Process (MDP) or Semi-Markov Decision Process (SMDP). A multi-step RL algorithm called Sarsa(,k) is proposed, which is a compromised variation of Sarsa and Sarsa(). It is equivalent to Sarsa if k is 1 and is equivalent to Sarsa() if k is infinite. Sarsa(,k) adjust its performance by setting k value. Two forms of Sarsa(,k), forward view Sarsa(,k) and backward view Sarsa(,k), are constructed and proved equivalent in off-line updating.

You might also be interested in these eBooks

Frontiers of Manufacturing and Design Science

View Preview

Info:

Periodical:

Applied Mechanics and Materials (Volumes 44-47)

Pages:

3611-3615

DOI:

https://doi.org/10.4028/www.scientific.net/AMM.44-47.3611

Citation:

Cite this paper

Online since:

December 2010

Authors:

Zhi Cong Zhang, Kai Shun Hu, Hui Yu Huang, Shuai Li, Shao Yong Zhao

Keywords:

Reinforcement Learning, Sarsa, Sarsa(λ), Sarsa(λ,k)

Export:

RIS, BibTeX

Price:

Permissions CCC:

Request Permissions

Permissions PLS:

Request Permissions

Сopyright:

Citation:

References

[1] S. Singh, T. Jaakkola, M. L. Littman, C. Szepesvári: Machine Learning Vol. 38(2000), pp.287-308.

DOI: 10.1023/a:1007678930559

Google Scholar

[2] K. Papadakia, V. Friderikos: Computers & Operations Research Vol. 35(2008), p.3848 – 3859.

Google Scholar

[3] D. Vengerov: Future Generation Computer Systems Vol. 25(2009), pp.728-736.

Google Scholar

[4] R. S. Sutton, A. G. Barto: Reinforcement Learning: An introduction. MIT Press, Cambridge, Massachusetts (1998).

Google Scholar

[5] G. A. Rummery, M. Niranjan: On-line Q-learning using connectionist systems. Technical Report CUED/F-INFENG/TR 166, Engineering Department, Cambridge University (1994).

Google Scholar

[6] G. A. Rummery: Problem Solving with Reinforcement Learning [Ph.D. dissertation]. Cambridge University (1995).

Google Scholar

[7] S. S. Singh, V. B. Tadić and A. Doucet: European Journal of Operational Research Vol. 178(2007), pp.808-818.

Google Scholar