Opportunistic target SINR setting for the MIMO broadcast channel

The sum capacity of the multiple input multiple output (MIMO) broadcast channel can be approached by a linear precoding technique called block diagonalization (BD) and by its recent extension that accounts for both in-cell and other-cell interference. In this paper we consider the problem of minimizing the transmitted power subject to a capacity constraint (the power control problem) and maximizing the sum information rate subject to a power constraint (the capacity problem) for the MIMO broadcast channel using BD precoding. Unlike previous works, we do not assume that a minimum signal-to-interference-plus-noise ratio (SINR) value is prescribed for each user but impose a fairness constraint on the user SINRs. Our approach preserves the degree of freedom of optimizing the SINR targets (in both problems) and thereby it allows to improve the system performance. We use the Augmented Lagrangian Penalty Function (ALPF) method to solve the respective constrained non-convex optimization problems and study the impact of different fairness constraints on the achievable maximum sum rate or minimum transmitted power. We find that BD with target SINR selection is a promising technique for the power control as well as the capacity problem.