Reducing the probability of network overload by spectrum trading

In this paper dynamic spectrum trading is addressed, where time-frequency resources are traded between network operators on a short-term basis to meet current demands. Specifically, we derive the probability of system overload for the entire network as well as for a single operator. Numerical results demonstrate that by dynamically adjusting the resource assignment to operators on a short-term basis, per-operator and system overload probabilities can be significantly reduced at given offered traffic.