Premium
Adaptive routing algorithm for network load balancing
Author(s) -
Une Hiroyuki,
Qian Fei,
Hirata Hironori
Publication year - 2011
Publication title -
ieej transactions on electrical and electronic engineering
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.254
H-Index - 30
eISSN - 1931-4981
pISSN - 1931-4973
DOI - 10.1002/tee.20680
Subject(s) - computer science , reinforcement learning , network packet , computer network , routing (electronic design automation) , equal cost multi path routing , source routing , routing algorithm , boundary (topology) , distributed computing , static routing , packet loss , routing table , algorithm , routing protocol , artificial intelligence , mathematics , mathematical analysis
This paper introduces a routing algorithm which distributes the network traffic for each feasible route based on the reinforcement learning scheme. The aim of this algorithm is to reduce the transfer delay caused by congestion of the network. Most routing algorithms based on the reinforcement learning scheme could not avoid forwarding traffic toward the source of data packets. This property creates transfer loops and a lot of packet loss. In particular, the transfer loop may cause change in the order of packets, and also retransmit on the TCP connection. Thus, we employ a method that estimates the lower boundary of the probability for valid route to the destination, and this lower boundary can be used to reject the routes which cause the transfer loop. © 2011 Institute of Electrical Engineers of Japan. Published by John Wiley & Sons, Inc.