Device-to-device (D2D)-enabled mobile edge computing (MEC) is an emerging technology which has been widely investigated in terrestrial networks. Different from most existing relevant work, where the network is base station-assisted and offloadings are made on homogeneous channels, this paper focuses on an unmanned aerial vehicle (UAV) swarm, where both the decentralized character and the heterogeneous data computation demands are considered. To fully utilize the limited time, spectrum and computation resources, the joint computation offloading and variable-width channel access problem is investigated. The problem is solved by a game-theoretic based solution. Specifically, the problem is first formulated into a game model which is proved to be an exact constrained potential game (ECPG). The game has at least one pure strategy generalized Nash equilibrium (GNE) and the best GNE is the global optimum of the problem. After that, to enable the UAV swarm reach the GNE autonomously, a distributed collective best response (COBR) algorithm is then proposed. The algorithm can converge to a GNE of the game, which is the local or global optimum of the proposed problem. Simulation results show that the proposed method can save about 10% energy than offloading on homogeneous channels.

This article focuses on the slot access problem for neighboring cooperation in unmanned aerial vehicle (UAV) swarms. To avoid the slot access process being hindered by unavailable topology information or information exchanges, a self-organized collision discovery mechanism is proposed. Each broadcaster can know whether its transmission is successful through the mechanism which provides the basic knowledge for finding the slot access strategy. Considering the distributed feature, the slot access problem is formulated as two game models. Both games are proved to have at least one Nash Equilibrum (NE) and the best NE is the optimum of the problem. Two distributed and synchronous algorithms are proposed to reach the NE. The first algorithm converges fast which satisfies the dynamic feature of UAV swarms and the second one converges to the optimum asymptotically. Moreover, to enhance the time efficiency of UAV swarms, the total number of required slots is investigated in some typical topologies and then conjectured to general ones. Simulation results verify that the proposed method is effective and the conjecture is true in almost all topologies.

In order to promote the spectral utilization, this article investigates the partially overlapping channel (POC) accessing problem in the wireless network. To reflect the heterogeneous characteristics of users, the optimization goal is set as maximizing the quality of service (QoE), instead of maximizing the throughput or minimizing the interference. The problem is formulated as a QoE maximization game and is then proved to be an ordinal potential game by utilizing the approximate relationship between interference and QoE. The proposed game is proved to have at least one pure Nash equilibrium (NE) and the best pure strategy NE point is an approximate global optimum of maximizing network QoE. A distributed algorithm is designed to reach the NE and it is proved that when the learning parameter is large enough, the algorithm asymptotically maximizes the network QoE. Simulation results verify the effectiveness of utilizing POCs and the proposed method.

Message interaction among neighborhood is necessary for unmanned aerial vehicles (UAVs) and Time Division Multiple Access is a feasible implementation by which each UAV broadcasts in one dedicated slot. However, the dynamic character requires the access process fast and the constrained energy limits the feedbacks during the process uninformative. Therefore, this article focuses on the slot access problem in UAV swarms. To be energy saving, a collision discovery method which is independent of valid information is designed to play the role of feedback. Furthermore, considering the decentralized structure, the slot access problem is formulated as a game model. To reach the Nash Equilibrium of the game fast, a synchronous and uncoupled learning algorithm is proposed to, in which all UAVs update simultaneously based on the information they obtain. The algorithm releases the requirement for a scheduler or a common control channel and is therefore applicable in UAV swarms. Simulation results verify the effectiveness of the proposed method and some investigations about the requirement for the number of slots are made.

In this paper, we investigate the hybrid access control in two-tier small cell networks from the perspective of incentive mechanism design, considering macro base station's private information. We formulate this problem as a Stackelberg game. To be specific, the macro base station (MBS) and small cell base stations (SBSs) are modeled as a leader with specific payoff preference and followers, respectively. A subsidy mechanism is adopted by MBS. Only when the SBS can provide acceptable service level for macro user (MUE), then MBS handovers MUE to SBS and provides subsidy to SBS; otherwise, MBS serves MUE by itself. We consider the impacts of MBS's private information on the final equilibrium of the proposed game. Theoretical analysis and simulation results show it is better for MBS to broadcast the private information to get more payoff from the perspective of incentive mechanism design.