|
|
For Full-Text PDF, please login, if you are a member of IEICE,
or go to Pay Per View on menu list, if you are a nonmember of IEICE.
|
Evolutionary Game Based Access Control Protocol in Wireless Networks with Mobile Routers
Ippei AOKI Koji YAMAMOTO Hidekazu MURATA Susumu YOSHIDA
Publication
IEICE TRANSACTIONS on Communications
Vol.E94-B
No.8
pp.2225-2234
Publication Date: 2011/08/01
Online ISSN: 1745-1345
DOI: 10.1587/transcom.E94.B.2225
Print ISSN: 0916-8516
Type of Manuscript: Special Section PAPER (Special Section on Deployment and Operation of New Internet Technology: Challenges and Approaches)
Category:
Keyword:
mobile router, evolutionary game theory, game theory, CSMA/CA,
Full Text: PDF(701.6KB)>>
Summary:
In existing systems of mobile routers, the frequency band is shared in uplinks from wireless terminals to mobile routers, and carrier sense multiple access with collision avoidance (CSMA/CA) is generally used as the medium access control protocol. To use the frequency band effectively, adaptive control is one promising approach. In this paper, a decentralized access control protocol in which mobile routers adaptively select the minimum contention window size is proposed. However, because of their mobility, which is one of the main difference between mobile routers and fixed access points, individual local area networks (LANs) consisting of the mobile routers and wireless terminals randomly interact with each other, and such random interactions can cause instability. To analyze the stability of the proposed control, evolutionary game theory is introduced because a system with random interactions between numerous decision-making entities can be analyzed by using evolutionary game theory. Using evolutionary game theory, the condition for existence of a convergence point is obtained. In addition, to implement the decentralized access control, a learning rule is proposed. In the proposed learning rule, each mobile router selects a strategy based on the result of past trials. From the simulation result, it is confirmed that the decentralized access control converges to a point closed to the stable state derived through evolutionary game theory.
|
|
