Optimal control of wireless networks with finite buffers

This paper considers network control for wireless networks with finite buffers. We investigate the performance of joint flow control, routing, and scheduling algorithms which achieve high network utility and deterministically bounded backlogs inside the network. Our algorithms guarantee that buffers...

Full description

Bibliographic Details
Main Authors: Le, Long Bao, Modiano, Eytan H., Shroff, Ness B.
Other Authors: Massachusetts Institute of Technology. Department of Aeronautics and Astronautics
Format: Article
Language:en_US
Published: Institute of Electrical and Electronics Engineers 2011
Online Access:http://hdl.handle.net/1721.1/66127
https://orcid.org/0000-0003-3577-6530
https://orcid.org/0000-0001-8238-8130
Description
Summary:This paper considers network control for wireless networks with finite buffers. We investigate the performance of joint flow control, routing, and scheduling algorithms which achieve high network utility and deterministically bounded backlogs inside the network. Our algorithms guarantee that buffers inside the network never overflow. We study the tradeoff between buffer size and network utility and show that if internal buffers have size (N - 1)/à ¿ then a high fraction of the maximum utility can be achieved, where à ¿ captures the loss in utility and N is the number of network nodes. The underlying scheduling/routing component of the considered control algorithms requires ingress queue length information (IQI) at all network nodes. However, we show that these algorithms can achieve the same utility performance with delayed ingress queue length information. Numerical results reveal that the considered algorithms achieve nearly optimal network utility with a significant reduction in queue backlog compared to the existing algorithm in the literature. Finally, we discuss extension of the algorithms to wireless networks with time-varying links.