On robust wireless network optimization using network criticality

Network criticality is a graph-theoretic metric that quantifies network robustness, and it was originally designed to capture the effect of environmental changes in core networks. This paper investigates the application of network criticality in designing robust power allocation and flow assignment algorithms for wireless networks. Achieving robust behavior in wireless networks is a challenging task due to constant changes in channel conditions and the interference. We consider network criticality as a natural robustness metric, and propose approaches to preserve the useful convexity properties of network criticality, while resolving issues related to the non-convexity of Shannon's capacity. The proposed optimization problems are applicable in all signalto-interference-plus-noise regimes.