Power efficient uplink scheduling in SC-FDMA: Bounding global optimality by column generation

We study resource allocation in cellular systems and consider the problem of finding a power efficient scheduling in an uplink single carrier frequency division multiple access (SC-FDMA) system with localized allocation of subcarriers, that is, the subcarriers allocated to a user equipment have to be consecutive in the frequency domain in each time slot. This problem is discrete and nonconvex, thus the use of suboptimal algorithms has been a common practice. We leverage the power of mathematical programming in order to approach global optimality or a tight bounding interval confining global optimum, to arrive at an effective scheme for gauging the performance of suboptimal algorithms. Toward this end, we first provide a straightforward integer linear programming formulation, and then an alternative and less trivial, so-called column-oriented, formulation. The latter is solved by column generation, which is a solution technique for large-scale optimization problems with certain characteristics. The computational evaluation demonstrates that the column generation method produces very high-quality subcarrier allocations that either coincide with the global optimum or enable an extremely sharp bounding interval. Hence the approach serves well for the purpose of benchmarking results for large-scale instances of power efficient SC-FDMA scheduling.