Intelligent Power Control Algorithm in Heterogeneous Wireless Cellular Networks

Appropriate power control algorithm can increase total network rate and reduce energy consumption through optimizing signal-to-interference-plus-noise ratio (SINR) received by users. Thus it is vital for heterogeneous cellular networks, which are composed of macro base stations (BSs) and micro BSs. However, most existing power control schemes neglect the dynamic traffic variation and backhaul bandwidth constraint due to the variability of wireless traffic and difficulty of problem solving. This results in the performance reduction and resource waste under the highly dynamic traffic. In view of this, we propose an intelligent power control algorithm based on accurate traffic prediction. First, hybrid prediction scheme, composed of long short term memory (LSTM) and historical average value (HAV), is used to predict the traffic demand of each BS. Based on the predicted results, we propose a power control scheme. It achieves a weighted tradeoff between the total network rate and energy consumption under the consideration of backhaul constraint. By using the properties of concave function and the function structure of backhaul constraint, the Marks-Wright algorithm and difference of two convex functions (DC) programming are respectively employed to solve the non-convex optimization model approximately and accurately. Simulation results demonstrate that the prediction accuracy of proposed scheme is 3.73% and 6.03% higher than that of the autoregressive integrated moving average (ARIMA) and support vector machine (SVM) algorithm, respectively. Compared with full power transmission, energy consumption of the proposed algorithm is reduced by 58.29%, while total network rate is increased by 3.26%.