A Cross Layer Optimization Model for Investigating the Impact of Partially Overlapping Channels on Wireless Mesh Networks Capacity

Wireless mesh networks have the particularity that both client and backhaul traffic often share the same radio channels. Moreover, available radio channels have to be shared not only between nodes of the same network, but also with potential adjacent ones. This may result in a high level of interference that has to be managed carefully. In this paper, we investigate by introducing a cross-layer optimization model based on the physical interference model, the gain of network capacity that can be achieved by exploiting the totality of available channels in the radio spectrum, as opposed to traditional operations limited to orthogonal channels. Both dynamic and static channel assignments are considered and also compared. We formulate the dynamic mode by a linear program and solve it through a column generation technique. As the static mode is more computationally challenging, we formulate it in the form of a mixed integer linear program and solve it through a branch and price technique. At the contrary to more optimistic works, we show through numerical experimentations that employing partially overlapping channels can lead to a moderate increase of the network capacity. We also show that the capacity obtained with a dynamic channel assignment does not necessarily result in a substantial improvement of capacity if an efficient static assignment is planned.