On protected traffic routing in wireless networks with partial multiple link failures

In this paper we study an optimization problem relevant for wireless networks. In such networks links are subject to capacity disruptions caused by changing weather/channel condition that force to adapt the modulation and coding schemes to the observed condition. To cope with this issue in network optimization, we need to consider a so called link availability set A and a routing strategy that can adapt to each state of link capacities corresponding to each point of A. We assume that A is finite and each alpha is an element of A specifies, for each link e is an element of epsilon fraction alpha(e) of its maximum capacity available in state a. For traffic protection we assume (possibly) over-dimensioned normal path-flows (i.e., path-flows defined for the availability state with all links fully available, the so called normal state) that are thinned in a failure state-dependent way to adapt to fluctuating reduced capacity. What is important, the normal flows cannot be increased or restored in any way. Moreover, we assume that the demand volumes to be realized in states alpha can be reduced as compared to the normal demands. To solve a link cost minimization problem corresponding to the above assumptions, we develop an original non-compact linear programming model together with its solution algorithm based on path generation. We also present a numerical study that illustrates the efficiency of the introduced model and compare the performance of the assumed flow-thinning routing with other routing strategies studied before. In the presented version, our model is applicable to wireless networks with non-interfering point-to-point links such as free-space optical links or microwave links.