Myopic versus clairvoyant admission policies in wireless networks

In this paper, we consider a dynamic scenario in which mobile users with elastic traffic arrive at a wireless heterogeneous system according to a Poisson arrival process. The wireless system consists of a set of overlapping network cells of different technologies. The protocols associated to each cell specify the throughput allocated to each user, given the cell load (i.e. the number of active users and their geographical positions). Meanwhile, mobile users (the players) can choose which cell to associate to. Suppose they collaborate to obtain an efficient and fair share of the global throughput. Then, two families of mechanisms can be considered: (i) either the users bargain at each stage of the system so as to obtain, given the present population, a fair share of the present throughput (repeated single-stage games), or (ii) users bargain so as to obtain a fair expected throughput during the total duration of their call (single stochastic game), taking into account the knowledge of the global (stochastic) arrival process and connection size. In this paper, we numerically compare the two policies and their resulting performance: does forecasting the future significantly impact the optimal mobile-to-cell association?

[1]  Dimitri P. Bertsekas,et al.  Dynamic Programming and Optimal Control, Two Volume Set , 1995 .

[2]  Hong-Shik Park,et al.  On achieving short-term QoS and long-term fairness in high speed networks , 2004, J. High Speed Networks.

[3]  Keith W. Ross,et al.  Uniformization for semi-Markov decision processes under stationary policies , 1987, Journal of Applied Probability.

[4]  Carl M. Harris,et al.  Fundamentals of queueing theory , 1975 .

[5]  Martin L. Puterman,et al.  Markov Decision Processes: Discrete Stochastic Dynamic Programming , 1994 .

[6]  Pierre Coucheney,et al.  Fair and Efficient User-Network Association Algorithm for Multi-Technology Wireless Networks , 2009, IEEE INFOCOM 2009.

[7]  Eitan Altman,et al.  User-Network Association in a WLAN-UMTS Hybrid Cell: Global & Individual Optimality , 2006, ArXiv.

[8]  Carl M. Harris,et al.  Fundamentals of queueing theory (2nd ed.). , 1985 .

[9]  Marceau Coupechoux,et al.  Network Controlled Joint Radio Resource Management for Heterogeneous Networks , 2008, VTC Spring 2008 - IEEE Vehicular Technology Conference.

[10]  Andreas Brandt,et al.  Waiting times for M/M systems under state-dependent processor sharing , 2008, Queueing Syst. Theory Appl..

[11]  J. Filar,et al.  Competitive Markov Decision Processes , 1996 .

[12]  Arnaud Legrand,et al.  Toward a fully decentralized algorithm for multiple bag-of-tasks application scheduling on grids , 2008, 2008 9th IEEE/ACM International Conference on Grid Computing.

[13]  L. Shapley,et al.  Stochastic Games* , 1953, Proceedings of the National Academy of Sciences.