Dynamic Spectrum Access for Multi-type Buyers in
Cognitive Radio Networks
Wei Zhong, Youyun Xu
Institute of Communica tions Engineering
PLAUST
Nanjing, China
Email: weizhong@ieee.org, yyxu@vip.sina.com
Jiaheng Wang
National Mobile Communications
Research Laborato ry
Southeast University
Nanjing, China
Email: jhwang@seu.edu.cn
Dapeng Li
College of Telecom.
and Information Engineering
Nanjing University of Posts and Telecom
Nanjing, China
Email: dapengli@nupt.edu.cn
Huaglory Tianfield
School of Engin eering a nd Buit Environment
Glasgow Caledonian University, U.K.
Email: H.Tianfield@gcu.ac.uk
Abstract—In this paper, we establish a game theoretic frame-
work of the auction based dynamic spectrum access for multi-
type buyers (i.e., the secondary users that have different risk
preferences) in cognitive radio networks where the auction
mechanisms of the primary user are parameterized and can be
adaptively designed. We design the utility functions of the multi -
type secondary users. The proposed game is a discrete game
having at least one mixed or pure strategy Nash equilibrium.
Based on the concept of learning automata, we propose a
distributed algorithm to learn the Nash equilibrium of the
proposed game with limited feedback information and prove
its convergence to the Nash equilibrium of our proposed game
with properly design ed utility function. Simulation results show
that our proposed algorithm is efficient and can achieve much
higher performance than th e traditional fixed auction mechanism
schemes and t he random algorithm.
I. INTRODUCTION
Cognitive radio is viewed as a novel approach for improving
the utilization of a precious natural resource: the radio electro-
magnetic spectrum [1]. Auction based spectr um management
provides a new way of spectrum utilization for the spectrum
markets in c ognitive radio networks and has attracted much
attention recently [2].
References [3]- [7] mainly focus on using game theory
to analyze the behaviors of secondar y users under a certain
auction mechanism. In particular, they assum e that the sec-
ondary users are symmetric and homogeneous. However, this
assumption may not be realistic in the practical cognitive radio
networks.
In practical cognitive networks, the secondary users are usu-
ally of different types and may have different risk preferences
in the auction, e.g., some secondary users may be regarded
as risk-seeking, who are rich or may want urgent message or
real-time service, some secon dary users may be regarde d as
risk-averse, who are poor or do not have urgent message and
real-time service , may be and some secondary users may be
regarded as risk-neu tral, who are intermediate between risk-
seeking and risk-averse.
References [8] [9] have shown that different auctio n mech-
anisms will result in different outcom es. This means a fixed
auction mechan ism may n ot be optimal for the auctioneer.
In order to obtain higher revenue, the auction mech anisms
should be adaptively designed by the auctioneer according to
the types of the secondary users. Therefore, the auction based
dynamic spectrum access with adaptive auction mechanism
design is essential and challenging in the cognitive radio
networks where the multi-type secondary users have different
risk preferences.
In this paper, we study the cognitive radio networks w here
multiple secondary users with different risk preferenc es (i.e.,
multi-type buyers) sense an idle c hannel and attempt to access
the channel through an auction simultaneously. Moreover, the
primary user, i.e., the au ctioneer, has multiple non-standard
sealed-bid auction mechanisms which can be adap tively de-
signed. Since it is difficult to let all terminals in the network
know the full information, then the questions are, how the
auctioneer should be designed the auction mechanism to
maximize the utility and how the secondary users should
choose their best bidding strategies with imperfect and limited
feedback information?
We apply game theory and learning automata theory [15] to
study this problem. Since the primary user and the secondary
users usually can not know the full information in cog nitive
networks, centralized algorithm is impractical in this situation.
Instead, we design a practical distributed algorithm to attain
the Nash equilibrium with only a little feedback information
and prove th e convergence of the proposed algorithm. Fur-
thermor e, the performance of our proposed algorithm is also
evalu ated.
Related works. Mechanism design [9] ha s been applied to
study the cognitive radio networks recently [10]. Reference
[10] considered two specific mechanisms to suppress cheating
and collusion behavior of selfish users; in contrast, we study
the pa rameterized auction mechanism design for multi-type
buyers. Adaptive auction mechanism design has been studied