Site-wide links

<<   October 2015   >>
Su Mo Tu We Th Fr Sa
w         1 2 3
w 4 5 6 7 8 9 10
w 11 12 13 14 15 16 17
w 18 19 20 21 22 23 24
w 25 26 27 28 29 30 31

  Search By:
Date Range:  
Start Date Click Here to open the calendar
End Date   Click Here to open the calendar
include date range

Event Details
Event Name: GCCIS PhD Colloquium Series
Category: Academic/College Events
Sub Category: GCCIS
Description: Adaptive Learning for Cross-layer Resource Management in Cognitive Radio Network

Presentation by
Wenbo Wang, Ph.D. Student


In recent years, cross-layer resource management has been widely considered as an efficient approach for improving the network performance. However, due to the factors such as limited knowledge about the wireless environment and spontaneous characteristics of device behaviors, designing an efficient distributed resource management scheme using the cross-layer paradigm becomes extremely difficult. In this talk, we propose a framework of adaptive learning for the resource management in the presence of stochastic and a priori unknown channel conditions. We formulate the optimization problem for resource management as a Multi-Agent Markov Decision Process (MAMDP), and solve it on-line using distributed reinforcement learning. The advantage of the proposed on-line learning method are that (i) it does not require a priori knowledge of the network details and (ii) it only needs the local action-state observation for updating the resource management policies. We present two cases of applying the proposed learning framework to the practical cross-layer resource management problems: (i) the joint PHY-APP layer resource allocation for real-time video transmission in distributed wireless networks and (ii) the cross-layer relay selection for multi-hop, multi-channel Cognitive Radio Networks (CRNs). Our simulation results show that the proposed learning algorithms achieve better performance compared to the conventional resource management schemes.
Schedules: 09/06/2013   (12:00 PM - 1:00 PM)

Contact: Joyce Hart
Phone: 475-6193
Cost: Free