Transmitted Power Formulation for the Optimization of Spectrum Aggregation in LTE-A over 800 MHz and 2 GHz Frequency Bands

This work starts by proposing a formulation to calculate the transmitter power needed to cover cells of different sizes, whilst maintaining the average signal to interference-plus-noise ratio constant, and near the maximum, for two Long Term Evolution (LTE) systems operating over non-contiguous frequency bands, 800 MHz and 2 GHz, with an integrated common radio resource management (iCRRM) entity. In the context of spectrum aggregation (SA), iCRRM is able to switch users between the two LTE-Advanced scenarios to facilitate the best user allocation and maximize the total network throughput in these LTE systems. We address a formulation based on the computation of the average received power and average co-channel interference in cellular topologies with frequency reuse pattern \(K\) = 3, keeping the presence of coverage holes insignificant, whilst considering the COST-231 Hata path loss model. We have verified how the normalized power increases as the cell radius increases. The objective of applying this formulation in the dimensioning process is to save power for the shortest coverage distances. It has been found that without SA the maximum average cell throughput is observed in the presence of 80 simultaneous users within the cell (40 for each LTE system, operating in different frequency bands). We have considered traced-based video sessions with a (video) bit rate of 128 kbps. In this scenario, through extensive simulations cell average supported throughput of approximately 6,800, 8,500 and 9,500 kbps have been obtained for the cases without SA (considering the sum of the 800 MHz and 2 GHz systems capacities), with a simple CRRM and with iCRRM, respectively. It was also found that when the peak throughput is achieved with 80 users, the average cell packet loss ratio without SA, with CRRM and iCRRM present values of 22, 11 and 7 %. The average cell delay with both CRRM and iCRRM entities is 22 ms, whereas without SA is equal to 32 ms. Finally, the cost/revenue tradeoff is analysed from the operator/service provider’s point of view, whose main goal is obtain the maximum profit from his business. It was found that CRRM increases the total profit in percentage, compared to a simple allocation, without SA. Nevertheless, the profit growth with iCRRM is even larger, from 253 to 296 % for \(R =\) 1,000 m and a price of 0.010 €/MByte. Therefore, our proposal for SA is convenient not only in terms of technical features and QoS, as loss and delay have been obtained within a range of reasonable values, but also in terms of economic aspects.