This article presents algorithms for the optimization of radio resources that can be implemented in dynamic mobile networks based on RoF. The algorithms are based on linear relaxation techniques for integer linear programming (ILP) problems. The formulation models a multi-tier structure of antennas with increasing radius. Considering this antenna structure, the optimizer performs dynamic cell merging and cell splitting to save resources and improve network availability. Both the integer and the relaxation formulations produced similar results for all the experiments, but the time of processing required by the relaxed algorithms was much shorter than those required by the integer algorithms for large instance of the problem, which highlights the advantages of relaxation techniques under time constraints.