The next generation networks seem to be too dense compared to the existing one, so a self-control mechanism, which determines the optimal cell size will be essential. In this paper, we present self organized cell size control algorithms, which maintain optimum system throughput and power consumption. Particularly, we investigate three different algorithms that control the cells size, while maintaining the optimum power consumption and block allocation in the network. These algorithms differ in terms of their decision area. The first one is based on a centralized control; the second one is a distributed approach; and the final one is based on a group distributed control. In order to evaluate their performance, these algorithms are tested upon two different simulation environment, which approach real scenarios. Our results indicate that the group distributed algorithm is the best approach for future network, since it has a good performance and about 10 times lower computational complexity when compared with the centralized approach.