A New Reinforcement Learning for Train Marshaling with Generalization Capability

Yoichi Hirashima

doi:10.4028/www.scientific.net/AMR.974.269

Paper Titles

Influence of Shear Rate on Proteins Separation, Molecular Weight Cut-Off and Average Pore Size of Polysulfone Blend Membranes
p.247

Characteristics of UV Irradiated Waste Biopolymer from Renewable Resources (Part 1)
p.252

Characteristics of UV Irradiated Waste Biopolymer from Renewable Resources (Part 2)
p.257

Surface Modified Nano Calcium Oxide for Base Heterogeneous Transesterification of Kappaphycus Alvarezii Seaweed to Biofuel
p.262

A New Reinforcement Learning for Train Marshaling with Generalization Capability
p.269

Performance Enhancement of 10 Gbps OCDMA Networks Using DPSK and DQPSK with Unique Code-Sequence
p.274

Vehicle Scheduling Model for Fresh Agriculture Products Pickup with Uncertain Demands
p.282

Application of Auxiliary Antenna Elements for SAR Reduction in the Human Head
p.288

Thinking on the Progressive Failure Analysis of the Slope
p.293

HomeAdvanced Materials ResearchAdvanced Materials Research Vol. 974A New Reinforcement Learning for Train Marshaling...

A New Reinforcement Learning for Train Marshaling with Generalization Capability

Abstract:

This paper proposes a new marshaling method for assembling an outgoing train. In the proposed method, each set of freight cars that have the same destination make a group, and the desirable group layout constitutes the best outgoing train. The incoming freight cars are classified into several ``sub-tracks'' searching better assignment in order to reduce the transfer distance of locomotive. Classifications and marshaling plans based on the transfer distance of a locomotive are obtained autonomously by a reinforcement learning system. Then, the number of sub-tracks utilized in the classification is determined by the learning system in order to yield generalization capability.

You might also be interested in these eBooks

View Preview

Info:

Periodical:

Advanced Materials Research (Volume 974)

Pages:

269-273

DOI:

https://doi.org/10.4028/www.scientific.net/AMR.974.269

Citation:

Cite this paper

Online since:

June 2014

Authors:

Yoichi Hirashima

Keywords:

Container Transfer Problem, Freight Train, Marshaling, Q-Learning, Scheduling

Export:

RIS, BibTeX

Price:

Permissions CCC:

Request Permissions

Permissions PLS:

Request Permissions

Сopyright:

Citation:

References

[1] C. F. Daganzo, R. G. Dowling, R. W. Hall: Railroad classification yard throughput: The case of multistage triangular sorting, Transportation Research, Part A, Vol. 2, No. 17 (1983), pp.95-106.

DOI: 10.1016/0191-2607(83)90063-8

Google Scholar

[2] U. Blasum, M. R. Bussieck, W. Hochstättler, C. Moll, H. -H. Scheel and T. Winter: Scheduling Trams in the Morning, Mathematical Methods of Operations Research, Vol. 49, No. 1 (2000), pp.137-148.

DOI: 10.1007/pl00020912

Google Scholar

[3] R. Jacob, P. Marton, J. Maue and M. Nunkesser: Multistage Methods for Freight Train Classification, Proc. 7th Workshop on Algorithmic Approaches for Transportation Modeling, Optimization, and Systems (2007) , pp.158-174.

DOI: 10.1002/net.20385

Google Scholar

[4] Y. Hirashima: An Intelligent Train Marshaling Based on the Processing Time Considering Group Layout of Freight Cars in IAENG Transactions on Electrical Engineering Volume 1, World Scientific (2013), pp.229-243.

DOI: 10.1142/9789814439084_0018

Google Scholar

[5] Y. Hirashima: A Reinforcement Learning System for Generating Train Marshaling Plan of Freight Cars Based on the Processing Time Considering Group Layout, Proc. Int. MultiConf. Engineers and Comp. Scientists (2013), pp.30-35.

DOI: 10.1142/9789814439084_0018

Google Scholar

[6] R. Sutton and A. Barto: Reinforcement Learning, MIT Press (1999).

Google Scholar