Game-Theoretic Approach to Capacity and Stability Evaluations of Decentralized Adaptive Route Selections in Wireless Ad Hoc Networks

Koji YAMAMOTO  Susumu YOSHIDA  

Publication
IEICE TRANSACTIONS on Communications   Vol.E88-B   No.3   pp.1009-1016
Publication Date: 2005/03/01
Online ISSN: 
DOI: 10.1093/ietcom/e88-b.3.1009
Print ISSN: 0916-8516
Type of Manuscript: Special Section PAPER (Special Section on Ubiquitous Networks)
Category: Network
Keyword: 
ad hoc networks,  decentralized control,  game theory,  network capacity,  routing,  

Full Text: PDF(428.3KB)>>
Buy this Article




Summary: 
A game-theoretic analysis is applied to the evaluation of capacity and stability of a wireless ad hoc network in which each source node independently chooses a route to the destination node so as to enhance throughput. First, the throughput of individual multihop transmission with rate adaptation is evaluated. Observations from this evaluation indicate that the optimal number of hops in terms of the achievable end-to-end throughput depends on the received signal-to-noise ratio. Next, the decentralized adaptive route selection problem in which each source node competes for resources over arbitrary topologies is defined as a game. Numerical results reveal that in some cases this game has no Nash equilibria; i.e., each rational source node cannot determine a unique route. The occurrence of such cases depends on both the transmit power and spatial arrangement of the nodes. Then, the obtained network throughput under the equilibrium conditions is compared to the capacity under centralized scheduling. Numerical results reveal that when the transmit power is low, decentralized adaptive route selection may attain throughput near the capacity.