Telecommun Syst (2013) 52:2003–2013 DOI 10.1007/s11235-011-9481-9 Evaluation of channel dependent bandwidth allocation in wireless access networks: centralized and distributed approach Stavroula Vassaki · Athanasios D. Panagopoulos · Philip Constantinou Published online: 15 June 2011 © Springer Science+Business Media, LLC 2011 Abstract The problem of bandwidth allocation in wireless access networks is studied in this paper, investigating the performance of two approaches. Firstly, we use centralized algorithms, such as bankruptcy division rules and Nash bar- gaining. Secondly, a distributed algorithm is proposed in or- der to find the optimal solution of the bandwidth allocation problem. In both approaches, the allocation rules are prop- erly modified to incorporate the influence of the channel state resulting in a more efficient and fair bandwidth allo- cation. The channel dependent centralized and distributed schemes are compared in terms of efficiency and fairness with a view to highlighting the advantages and disadvan- tages of every approach. Keywords Resource allocation · Bankruptcy problem · Auction · Design mechanism 1 Introduction Resource, and specifically bandwidth, allocation in wire- less access networks consists one of the classical problems, which has been widely addressed by the research commu- nity [1]. Since bandwidth determines the maximum trans- mission rate, given from the well-known Shannon capac- ity, and puts a limit on the channel access rate, its optimum S. Vassaki ( ) · A.D. Panagopoulos · P. Constantinou National Technical University of Athens (NTUA), Athens, Greece e-mail: voula@mobile.ntua.gr A.D. Panagopoulos e-mail: thpanag@mobile.ntua.gr P. Constantinou e-mail: fkonst@mobile.ntua.gr management remains a great challenge. More interesting de- sign problems arise in 4G wireless networks, where hetero- geneous wireless technologies (WLAN, cellular, WMAN, satellite etc.) are employed in order to provide high data rate services and to guarantee the Quality of Service (QoS). In this paper, the problem of bandwidth sharing to users’ terminals with different channel state is studied using two different approaches. In the first approach, we assume that each terminal suffers from different attenuation due to chan- nel conditions and also it runs a different application de- manding different data rates. The bandwidth allocation problem is solved in a centralized way using basic division rules of bankruptcy problems [2]. Additionally, in order to be able to compare the outcome of the division rules, the same problem is also analyzed using principles from cooper- ative game theory and specifically applying the well-known Nash Bargaining Solution (NBS). Most of this approach has been presented in [3]. However, in this paper, the Random Arrival rule has been replaced by another rule with less com- putational complexity. In the second approach, we consider a more realistic sce- nario, where each user has a different valuation function for the allocated bandwidth depending on the type of his appli- cation. In order to achieve an allocation that maximizes the sum of utilities of all the users, we apply the VCG-Kelly mechanism [4] with an adjustment so as to incorporate the influence of the channel state. The paper is organized as follows: in Sect. 2, prior work on game-theoretic models and design mechanisms for band- width allocation are briefly presented. The system model un- der consideration and the problem of bandwidth allocation are described in Sect. 3. In Sect. 4, the centralized approach is analyzed in order to find the optimal resource allocation. Specifically, the bankruptcy allocation problem is explained and the basic division rules are briefly presented. Moreover,