| Article ID: | iaor1989927 |
| Country: | Japan |
| Volume: | J72-D-I |
| Issue: | 9 |
| Start Page Number: | 669 |
| End Page Number: | 677 |
| Publication Date: | Sep 1989 |
| Journal: | Transactions of the Institute of Electronics, Information and Communication Engineers |
| Authors: | Qian Fei, Hirata Hironori, Fukao Takeshi |
| Keywords: | communications, engineering, learning |
The routing problem is one of the most important problems to increase the performance of packet switched computer networks. In this paper, two new distributed quasi-static routing algorithms are proposed. The learning automaton located at each node of the network selects routing choices and modifies its strategy under the network conditions. Simulation studies demonstrate the feasibility of the methodology for learning automaton routing, and show the usefulness of the proposed algorithms for the networks with large number of nodes or heavy traffic. [In Japanese.]