Date Added: Nov 2012
The authors consider wireless mesh networks in which the nodes are utilizing cognitive radios and try to opportunistically gain access to spectrum resources. In such networks, the timely delivery of the traffic is a challenging task due to the licensed (primary) users' activities and their traffic characteristics. To overcome this challenge they propose an algorithm that minimizes the end-to-end delay through joint routing and spectrum resources allocation. The network is analyzed from a queuing theory perspective to capture the effects of dynamic spectrum availability on mesh network's traffic. The joint routing and resource allocation problem is formulated as a non-linear integer programming problem, for which they propose a decentralized solution based on the Lagrangian dual problem.