Value of information in optimal flow-level scheduling of users with Markovian time-varying channels

In this paper, we design, characterize in closed-form, and evaluate a new index rule for Markovian time-varying channels, which gives rise to a simple opportunistic scheduling rule for flow-level scheduling in wireless downlink systems. For user channels, we employ the Gilbert-Elliot model with a flow-level interpretation: the channel condition follows a general two-state Markov chain with distinct probabilities of finishing the flow transmission. The index value of the bad channel condition takes into account both the one-period and the steady-state potential improvement of the service completion probability, while the good channel condition gets an absolute priority with the [email protected] (well-known to be throughput-optimal) as the tie-breaking rule. Our computational study confirms near-optimality of the proposed rule in most of the instances, and suggests that information about the channels steady state is often enough to achieve near-optimality.

[1]  Urtzi Ayesta,et al.  Scheduling in a Random Environment: Stability and Asymptotic Optimality , 2011, IEEE/ACM Transactions on Networking.

[2]  Matthew S. Grob,et al.  CDMA/HDR: a bandwidth-efficient high-speed wireless data service for nomadic users , 2000, IEEE Commun. Mag..

[3]  R. Srikant,et al.  Throughput-Optimal Opportunistic Scheduling in the Presence of Flow-Level Dynamics , 2011, IEEE/ACM Transactions on Networking.

[4]  Christian M. Ernst,et al.  Multi-armed Bandit Allocation Indices , 1989 .

[5]  Sem C. Borst,et al.  Instability of MaxWeight Scheduling Algorithms , 2009, IEEE INFOCOM 2009.

[6]  P. Whittle.  Restless bandits: activity allocation in a changing world , 1988, Journal of Applied Probability.

[7]  N.B. Shroff,et al.  Optimal opportunistic scheduling in wireless networks , 2003, 2003 IEEE 58th Vehicular Technology Conference. VTC 2003-Fall (IEEE Cat. No.03CH37484).

[8]  P. Jacko.  Marginal productivity index policies for dynamic priority allocation in restless bandit models , 2011 .

[9]  P. Whittle.  Restless Bandits: Activity Allocation in a Changing World , 1988 .

[10]  Thomas Bonald.  A Score-Based Opportunistic Scheduler for Fading Radio Channels , 2004 .

[11]  Samuli Aalto,et al.  Flow-level stability and performance of channel-aware priority-based schedulers , 2010, 6th EURO-NGI Conference on Next Generation Internet.

[12]  R. Weber,et al.  On an index policy for restless bandits , 1990, Journal of Applied Probability.

[13]  Leandros Tassiulas,et al.  Dynamic server allocation to parallel queues with randomly varying connectivity , 1993, IEEE Trans. Inf. Theory.

[14]  Jean Walrand,et al.  The c# rule revisited , 1985 .

[15]  Aleksi Penttinen,et al.  On the optimal trade-off between SRPT and opportunistic scheduling , 2011, SIGMETRICS '11.

[16]  J. Bather,et al.  Multi‐Armed Bandit Allocation Indices , 1990 .

[17]  Sem C. Borst,et al.  Flow-level performance and capacity of wireless networks with user mobility , 2009, Queueing Syst. Theory Appl..

[18]  Raymond Knopp,et al.  Information capacity and power control in single-cell multiuser communications , 1995, Proceedings IEEE International Conference on Communications ICC '95.

[19]  J. Walrand,et al.  The cμ rule revisited , 1985, Advances in Applied Probability.

[20]  Philip A. Whiting,et al.  Convergence of proportional-fair sharing algorithms under general conditions , 2004, IEEE Transactions on Wireless Communications.

[21]  Urtzi Ayesta,et al.  A modeling framework for optimizing the flow-level scheduling with time-varying channels , 2010, Perform. Evaluation.

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

[23]  Gustavo de Veciana,et al.  Throughput optimality of delay-driven MaxWeight scheduler for a wireless system with flow dynamics , 2009 .

[24]  José Niño-Mora,et al.  Dynamic priority allocation via restless bandit marginal productivity indices , 2007, 2304.06115.

[25]  E. Gilbert.  Capacity of a burst-noise channel , 1960 .

[26]  G. Veciana,et al.  Throughput optimality of delay-driven MaxWeight scheduler for a wireless system with flow dynamics , 2009, 2009 47th Annual Allerton Conference on Communication, Control, and Computing (Allerton).

[27]  José Niño Mora.  Restless Bandits, Partial Conservation Laws and Indexability , 2000 .

[28]  Peter Jacko.  Adaptive Greedy Rules for Dynamic and Stochastic Resource Capacity Allocation Problems , 2010 .

[29]  Demosthenis Teneketzis,et al.  ON THE OPTIMALITY OF AN INDEX RULE IN MULTICHANNEL ALLOCATION FOR SINGLE-HOP MOBILE NETWORKS WITH MULTIPLE SERVICE CLASSES , 2000 .

[30]  Sem C. Borst.  User-level performance of channel-aware scheduling algorithms in wireless data networks , 2005, IEEE/ACM Transactions on Networking.

[31]  José Niño-Mora.  Characterization and computation of restless bandit marginal productivity indices , 2007, VALUETOOLS.

[32]  Gustavo de Veciana,et al.  Balancing SRPT prioritization vs opportunistic gain in wireless systems with flow dynamics , 2010, 2010 22nd International Teletraffic Congress (lTC 22).