Wireless powered device-to-device (D2D) communications can empower D2D communications via a power station (PS) in a self-sustainable manner. In this paper, aided by an incentive mechanism, we investigate the dynamic resource allocation for content transmission in wireless powered D2D communications with self-interested nodes including D2D transmitters (D2D-Txs) and PS. Stochastic optimization is developed to maximize the average network utility under the constraints of the limited data buffer, energy capacity, and incentives. Then, to coordinate the different timescales for the network state variation, the problem of stochastic optimization is converted into two subproblems via two-timescale Lyapunov optimization technique. Specifically, at the large timescale where the D2D-Tx's resource states (data queue, energy resource, and incentive) change slowly, we obtain the solution of a joint rate adaption and energy trading problem. While, at the small timescale where the channel state experiences rapid variation, we develop a closed-form expression of power factor by solving a transmit power control problem, which is non-convex owing to the interference among D2D-Txs. Additionally, an online two-timescale resource allocation (OTTRA) algorithm is proposed, and the performance bounds of the algorithm are characterized theoretically in terms of the utility-delay tradeoff. The numerical results exhibit that the OTTRA algorithm not only encourages the cooperative content transmission and ET among self-interested nodes, but also ensures a satisfied network performance in the long term.INDEX TERMS Wireless powered D2D communications, two timescales, self-interested nodes.
I. INTRODUCTIONWith the upsurge growth of data traffic and the increasing number of smart phones, device-to-device (D2D) communications have recently been widely studied as one of the key technologies to realize the 5G communication networks. By taking advantage of cooperative transmitters, D2D communications can improve the network performance (e.g., transmission capacity, delay, etc.) through direct communications between D2D devices without traversing the base station (BS) [1]. In such D2D networks, most D2D transmitters (D2D-Txs) are powered by limited battery capacity, and shall be recharged or replaced when their energy resource is depleted, which will hinder the fruitful development of D2D communication networks. Moreover, the devices in practical D2D communications are always endowed with smart autonomic functions, which are natural self-interested nodes [2]. These self-interested nodes just care about their own profits and will present selfishness, i.e., be unwilling to participate in cooperative content transmission. In D2D communication system, there are many kinds of selfish behaviors for the self-interested nodes, e.g., the hesitating caching for different social ties [3], unfaithful reporting for privacy protection and so on. The most common is that the D2D-Tx as an individual may refuse to forward the data for others to save their finite res...