## Power Allocation and Routing in Multibeam Satellites With Time-Varying Channels (2003)

### Cached

### Download Links

- [lids.mit.edu]
- [www.mit.edu]
- [www-bcf.usc.edu]
- [www-bcf.usc.edu]
- DBLP

### Other Repositories/Bibliography

Venue: | IEEE TRANSACTIONS ON NETWORKING |

Citations: | 67 - 16 self |

### BibTeX

@ARTICLE{Neely03powerallocation,

author = {Michael J. Neely and Eytan Modiano and Charles E. Rohrs},

title = {Power Allocation and Routing in Multibeam Satellites With Time-Varying Channels},

journal = {IEEE TRANSACTIONS ON NETWORKING},

year = {2003},

volume = {11},

pages = {138--152}

}

### Years of Citing Articles

### OpenURL

### Abstract

We consider power and server allocation in a multibeam satellite downlink which transmits data to different ground locations over time-varying channels. Packets destined for each ground location are stored in separate queues and the server rate for each queue depends on the power ( ) allocated to that server and the channel state ( ) according to a concave rate-power curve ( ). We establish the capacity region of all arrival rate vectors ( 1 ... ) which admit a stabilizable system. We then develop a power-allocation policy which stabilizes the system whenever the rate vector lies within the capacity region. Such stability is guaranteed even if the channel model and the specific arrival rates are unknown. Furthermore, the algorithm is shown to be robust to arbitrary variations in the input rates and a bound on average delay is established. As a special case, this analysis verifies stability and provides a performance bound for the Choose-the- -Largest-Connected-Queues policy when channels can be in one of two states (ON or OFF) and servers are allocated at every timestep ( ). These results are extended to treat a joint problem of routing and power allocation in a system with multiple users and satellites and a throughput maximizing algorithm for this joint problem is constructed. Finally, we address the issue of interchannel interference and develop a modified policy when power vectors are constrained to feasible activation sets. Our analysis and problem formulation is also applicable to power control for wireless systems.

### Citations

753 |
Nonlinear Programming
- Bertsekas
- 1995
(Show Context)
Citation Context ...hed work and channel-state vectors and , the problem (16) becomes a standard concave maximization problem: Maximize , subject to the simplex constraint . Using standard Lagrange multiplier techniques =-=[22]-=-, it can be shown that a solution is optimal if and only if power is allocated according to the constraints so that scaled derivatives are equalized to some value for all queues which receive nonzero ... |

530 |
Applied Probability and Queues
- Asmussen
- 1987
(Show Context)
Citation Context ...nished work in the queue exceeds the value . Stability in this case is identical to the usual notion of stability defined in terms of a vanishing complementary occupancy distribution [3], [10], [13], =-=[17]-=-. Lemma 1 (Queue Stability): For a single-queue system with general input and server rate processes and , a necessary condition for stability is . If the arrival process and the rate process evolve ac... |

394 | Achieving 100% Throughput in an Input-Queued Switch
- McKeown, Mekkittikul, et al.
- 1999
(Show Context)
Citation Context ...4], similar analysis is used to treat server allocation in a network with time-varying connectivity. Such a technique for proving stability has also been used in the switching literature [9]–[12]. I=-=n [10]-=-, an packet switch with blocking is treated and input/output matching strategies are developed to ensure 100% throughput whenever arrival rates are within the capacity region. In [12] and [13], the me... |

317 |
Capacity of a burst-noise channel
- Gilbert
- 1960
(Show Context)
Citation Context ...dium, and Bad channel conditions, with transition probabilities shown in Fig. 8. Such a three-state system has been considered in [24] and extends the well-known two-state Gilbert–Elliott model [28]=-=, [29]-=- for satellite and wireless channels. In each state, we assume signal attenuation is log-normally distributed with a given mean and variance. Such a distribution is consistent with the Karasawa model ... |

287 | Input versus output queueing on a space division switch - Karol, Hluchyj, et al. - 1987 |

184 | Dynamic server allocation to parallel queues with randomly varying connectivity
- Tassiulas, Ephremides
- 1993
(Show Context)
Citation Context ...LY et al.: POWER ALLOCATION AND ROUTING IN MULTIBEAM SATELLITES 139 Previous work on queue control problems for satellite and wireless applications is found in [1]–[8], [14], [15], [18], and [20]. I=-=n [2]-=-, a parallel queue system with a single server is examined, where every timeslot the transmit channels of the queues vary between ON and OFF states and the server selects a queue to service from those... |

165 | Discrete Stochastic Processes - Gallager - 1996 |

117 |
Estimates of error rates for codes on burst-noise channels
- Elliott
- 1963
(Show Context)
Citation Context ...od, Medium, and Bad channel conditions, with transition probabilities shown in Fig. 8. Such a three-state system has been considered in [24] and extends the well-known two-state Gilbert–Elliott mode=-=l [28]-=-, [29] for satellite and wireless channels. In each state, we assume signal attenuation is log-normally distributed with a given mean and variance. Such a distribution is consistent with the Karasawa ... |

103 | Gamal, “Energy-efficient packet transmission over a wireless link
- Uysal-Biyikoglu, Prabhakar, et al.
- 2002
(Show Context)
Citation Context ...on results in Section VII).sNEELY et al.: POWER ALLOCATION AND ROUTING IN MULTIBEAM SATELLITES 139 Previous work on queue control problems for satellite and wireless applications is found in [1]–[8]=-=, [14]-=-, [15], [18], and [20]. In [2], a parallel queue system with a single server is examined, where every timeslot the transmit channels of the queues vary between ON and OFF states and the server selects... |

96 |
A Practical Scheduling Algorithm to Achieve 100% Throughput
- Mekkittikul, McKeown
- 1998
(Show Context)
Citation Context ...traints. In [4], similar analysis is used to treat server allocation in a network with time-varying connectivity. Such a technique for proving stability has also been used in the switching literature =-=[9]��-=-�[12]. In [10], an packet switch with blocking is treated and input/output matching strategies are developed to ensure 100% throughput whenever arrival rates are within the capacity region. In [12] an... |

70 |
Dynamic power allocation and routing for satellite and wireless networks with time varying channels
- Neely
- 2003
(Show Context)
Citation Context ...to bandwidth limitations and develop a modified policy when power vectors are constrained to activation sets. This analysis makes use of a Lyapunov function defined over the state of the queues. 1 In =-=[20]-=-, we show the same policy is stabilizing for Markov modulated input and channel dynamics (see simulation results in Section VII).sNEELY et al.: POWER ALLOCATION AND ROUTING IN MULTIBEAM SATELLITES 139... |

66 | Optimal Control of Two Interacting Service Stations - Hajek - 1984 |

57 |
Scheduling and performance limits of networks with constantly changing topology
- Tassiulas
- 1997
(Show Context)
Citation Context ...ding to memoryless processes and have exponentially distributed length. A Lyapunov function is used to establish a stabilizing routing and scheduling policy under network connectivity constraints. In =-=[4], -=-similar analysis is used to treat server allocation in a network with time-varying connectivity. Such a technique for proving stability has also been used in the switching literature [9]–[12]. In [1... |

47 | Duality and linear programs for stability and performance analysis queueing networks and scheduling policies
- Kumar, Meyn
- 1996
(Show Context)
Citation Context ...12]. In [10], an packet switch with blocking is treated and input/output matching strategies are developed to ensure 100% throughput whenever arrival rates are within the capacity region. In [12] and =-=[13]-=-, the method of Lyapunov stability analysis is used to prove that queues are not only stable but have finite backlog moments. The main contribution in this paper is the formulation of a general power-... |

45 | Power and server allocation in a multi-beam satellite with time varying channels
- Neely, Modiano, et al.
- 2002
(Show Context)
Citation Context ... simulation results in Section VII).sNEELY et al.: POWER ALLOCATION AND ROUTING IN MULTIBEAM SATELLITES 139 Previous work on queue control problems for satellite and wireless applications is found in =-=[1]��-=-�[8], [14], [15], [18], and [20]. In [2], a parallel queue system with a single server is examined, where every timeslot the transmit channels of the queues vary between ON and OFF states and the serv... |

19 | Bounds on average delays and queue size averages and variances in input queued cell-based switches
- LEONARDI, MELLIA, et al.
(Show Context)
Citation Context ...nts. In [4], similar analysis is used to treat server allocation in a network with time-varying connectivity. Such a technique for proving stability has also been used in the switching literature [9]�=-=��[12]-=-. In [10], an packet switch with blocking is treated and input/output matching strategies are developed to ensure 100% throughput whenever arrival rates are within the capacity region. In [12] and [13... |

15 |
Large Deviations in High Speed Communication Networks
- Paschalidis
- 1996
(Show Context)
Citation Context ... ergodic finite-state Markov chain, then a sufficient condition for stability is . Proof: The sufficient condition for Markovian arrivals and linespeeds is well known (see large deviations results in =-=[16]-=-). The necessary condition is proven in the Appendix by showing that if , there exist arbitrarily large times such that the average fraction of time that the unfinished work is above during is greater... |

14 | Scheduling with asynchronous service opportunities with applications to multiple satellite systems - Carr, Hajek - 1993 |

12 | Some capacity results for ad-hoc networks
- Toumpis, Goldsmith
- 2000
(Show Context)
Citation Context ...ults in Section VII).sNEELY et al.: POWER ALLOCATION AND ROUTING IN MULTIBEAM SATELLITES 139 Previous work on queue control problems for satellite and wireless applications is found in [1]–[8], [14]=-=, [15]-=-, [18], and [20]. In [2], a parallel queue system with a single server is examined, where every timeslot the transmit channels of the queues vary between ON and OFF states and the server selects a que... |

9 | Extension of the Optimality of the Threshold Policy in Heterogeneous Multiserver Queueing Systems - Viniotis, Ephremides - 1988 |

9 |
Dynamic routing to parallel time-varying queues with applications to satellite and wireless networks
- Neely, Modiano, et al.
- 2002
(Show Context)
Citation Context ...n Section VII).sNEELY et al.: POWER ALLOCATION AND ROUTING IN MULTIBEAM SATELLITES 139 Previous work on queue control problems for satellite and wireless applications is found in [1]–[8], [14], [15]=-=, [18]-=-, and [20]. In [2], a parallel queue system with a single server is examined, where every timeslot the transmit channels of the queues vary between ON and OFF states and the server selects a queue to ... |

8 |
A general multibeam satellite switching algorithm
- Bongiovanni, Tang, et al.
- 1981
(Show Context)
Citation Context ...ulation results in Section VII).sNEELY et al.: POWER ALLOCATION AND ROUTING IN MULTIBEAM SATELLITES 139 Previous work on queue control problems for satellite and wireless applications is found in [1]�=-=��[8]-=-, [14], [15], [18], and [20]. In [2], a parallel queue system with a single server is examined, where every timeslot the transmit channels of the queues vary between ON and OFF states and the server s... |

7 |
Alnett. A new prediction method for tropospheric scintillation on earth-space paths
- Karasawa, Yamada, et al.
- 1988
(Show Context)
Citation Context ...ng experiments show that channel states could be modeled as i.i.d. during clear weather conditions (due to the observed rapid fluctuation of signal attenuation from scintillations in the Ka band [23]�=-=��[27]-=-). However, in rainy weather, future channel states are highly dependent on the current state. In [23] and [24], it is shown that the channel-state variations can be modeled as a Markov process. In [2... |

6 |
Predicting and adapting satellite channels with weatherinduced impairments
- Choi, Chan
- 2002
(Show Context)
Citation Context ... selects vertex points at every timeslot if power curves are piecewise linear. This power-allocation formulation generalizes a simpler problem of server allocation. Assume that there are servers 2 In =-=[23]-=- and [24], it is shown that satellite channel states can be accurately predicted up to one second into the future, where attenuation levels in the Ka band are estimated to within 61 db of accuracy in ... |

6 | Channel characterization and modelling for ka-band very small aperture terminals - Alouini, Borgsmiller, et al. - 1997 |

5 | Convexity and optimal load distributions in work conserving ∗/∗/1 queues - Neely, Modiano - 2001 |

5 |
Channel prediction and adaptation over satellite channels with weatherinduced impairments
- Choi
- 2000
(Show Context)
Citation Context ...vertex points at every timeslot if power curves are piecewise linear. This power-allocation formulation generalizes a simpler problem of server allocation. Assume that there are servers 2 In [23] and =-=[24]-=-, it is shown that satellite channel states can be accurately predicted up to one second into the future, where attenuation levels in the Ka band are estimated to within 61 db of accuracy in both clea... |

3 | Ka-band scintillations: Measurements and model predictions - Mayer, Jaeger, et al. - 1997 |