Joint Base Station Association and Power Allocation for uplink sum-rate maximization

In this paper, the problem of sum-rate maximization with Quality of Service (QoS) for a multi-cell multi-user uplink is addressed. The problem is formulated as a Mixed Integer Nonlinear Programming (MINLP) problem with non-convex feasible region and hence is difficult to solve. A primal-dual infeasible-interior-point method (IIPM) is applied to jointly optimize resources. With this method, simultaneously optimizing the Base Station Association (BSA) and Power Allocation (PA) is possible. Further to reduce the size, the problem is decomposed into two subproblems. The NP-hard Integer Programming (IP) BSA subproblem from the decomposition is solved by two different methods. One method uses the IIPM and other uses a Semidefinite Programming formulation. The PA subproblem is solved iteratively by IIPM. Simulation results converge to the optimum obtained by an existing exhaustive search. Apart from the sum-rate objective, the IIPM is applicable to broad class of utility functions and objectives and it also eliminates the requirement of an initial primal feasible point to begin the algorithm.