Stochastic approximate dynamic programming with link estimation for high quality path selection in Wireless Mesh Networks

A lot of work has recently been published regarding metrics that could identify high quality paths in Wireless Mesh Networks (WMN). While results are encouraging, no optimal strategy has yet been identified that could estimate link quality and incorporate both the link reliability measurements as well as the bandwidth capacity. Furthermore, link estimation remains an open problem. Considering multi-user environment, any optimal solution would also need to consider multiple communication flows. These arguments have led us to study an approximate dynamic programming (DP) solution capable of utilizing limited network knowledge and stochastic process. Instead of proposing yet another link metric, we analyze a stochastic DP solution to the routing problem using a well established routing metric. Unlike deterministic DP where communication demands are fixed a priori and an optimal path is calculated before any real demands are known, we consider a more realistic scenario with stochastic metrics and formulate the optimal strategy for routing in WMNs. Performance results are given using simulation results.