Accurate optimization models for interference constrained bandwidth allocation in cellular networks

Abstract

In cellular networks, the signal-to-interference-plus-noise ratio (SINR) is a key metric for link availability and quality. For network planning purposes, a straightforward modeling unfortunately yields numerically difficult optimization models. Further, given a required data rate of a link, its bandwidth consumption depends nonlinearly on the SINR.In this paper, we develop two novel approaches to jointly model SINR-based link availability and bandwidth requirements accurately. The first approach is a set-wise formulation from a user’s point of view, while the second one exploits discrete channel quality indicators. We compare these formulations with three known approximate approaches numerically, revealing the clear outperformance of our approaches in terms of exactness. Moreover, since the exact models comprise an exponential number of either variables or constraints, we discuss their pros and cons in a further computational study and develop a more efficient algorithm dealing implicitly with the involved constraints.