Application of the Water Filling Algorithm to the Sum Rate Problem with Minimum Rate and Power Constraint

The problem of weighted throughput maximization is generally non convex and computationally intense. A solution, known as MLFP-based power allocation or MAPEL algorithm, transforms the problem into a multiplicative linear fractional programming problem and then uses a method based on polyblock outer approximation. In this paper, we are going to use some techniques for reducing MAPEL computational time, in particular regarding the minimum rate constraint in the optimization problem. Furthermore, we are introducing, in the optimization procedure, the water filling principle in order to respect a maximum available power per user when the allocation has to be performed in more resource blocks.