Optimizing resource allocation for D2D communications with incomplete CSI

Device-to-device (D2D) communication is a new technology in cellular networks which enables direct communications between nearby devices. It could reuse the uplink spectrum of the cellular users (CUs) thus will introduce interference. Efficient resource allocation methods can coordinate interference and further improve network capacity. The difficulty of obtaining perfect channel state information (CSI) makes the resource allocation more complex. By using the statistical information of the unknown channel fading factors, we propose an optimal power control algorithm for any CU and any D2D pair sharing the same channel. Based on that, we formulate the resource allocation problem for multiple CUs and multiple D2D pairs with incomplete CSI as a maximum weight matching problem. We then solve the problem by combining the optimal power control algorithm and the maximum weight matching algorithm. Simulation results demonstrate the efficiency of our algorithms in achieving high D2D rates while guaranteeing the minimum rates required by CUs.