Search

Mobile edge computing has been an important supplement for traditional cloud computing architecture to offer low-delay computing services to mobile users. However, it is in general impossible for edge service providers to overdeploy so much edge resources to satisfy the rapidly increasing while diverse user demands. In this paper, we study a mobile edge computing system consisting of a service platform, cloudlets joining the system, and mobile users. In this study, we focus on a profit-driven perspective such that the service platform purchases computation resource from the resource-rich cloudlets and makes profit by processing tasks from user side. The design objective is to maximize the platform profit subject to budget constraint and stringent delay requirements for task processing. We formulate this problem as a mixed integer programming problem. Due to the NP-hardness of the problem, we design a logic based Benders decomposition algorithm as the offline solution. We further study the scenario where the task arrivals from user side and resource availability at the cloudlets are both stochastic and unknown in advance. We accordingly propose a Multi-Armed Bandit learning based resource purchasing and greedy task scheduling algorithm for the online scenario. Simulations results show the high performance of our proposed algorithms.

In wireless multi-hop networks, a fundamental problem is to disseminate continuous data traffic from a source node to all other network nodes, which is known as the broadcast problem. Such a problem becomes even more complicated in wireless multi-hop networks with energy-harvesting capabilities at nodes when facing the interaction between stochastics of traffic arrivals at the source and randomness of energy-harvesting process at nodes. In such networks, the energy consumable at a node cannot exceed the amount of the energy harvested at the node. In this paper, we investigate the throughput-optimal broadcast problem in time-varying directed acyclic wireless multi-hop networks with such energy harvesting constraints. The topologies of such networks change dynamically with time while satisfying the directed acyclic property and the energy arrival time and harvested amount at a node are random, which causes the consumable energy in each time slot to fluctuate with time. Existing throughput-optimal broadcast algorithms did not consider such energy-harvesting constraints in their designs and therefore their throughput-optimal properties do not hold anymore in such a network. In this paper, we characterize the energy-harvesting uncertainties at nodes by using time-varying per-slot-based supportable transmission rates of wireless links. We consider the time-varying property of supportable link transmission rates caused by energy-harvesting dynamics in the per-slot transmission scheduling and propose an online max-weight broadcast algorithm. We derive a tight upper bound of broadcast capacity of the wireless networks under study in this paper and further prove that our proposed algorithm is throughput-optimal. We evaluate the throughput and latency performance of the proposed algorithm by simulations and the simulation results affirm our theoretical analysis.

The problem of disseminating continuous data flow from a given source node to all other network nodes is known as the dynamic broadcast problem, which is also a fundamental problem in multi-hop wireless networks (e.g., wireless ad hoc and sensor networks). Due to link unreliability and node mobility, the topology of a network typically changes with time. Little work has been conducted to address the dynamic broadcast problem in time-varying wireless multi-hop networks, especially under the popular Signal-to-Interference-plus-Noise-Ratio (SINR) interference model. In this paper, we study the dynamic broadcast problem in SINR-based time-varying directed acyclic multi-hop wireless networks formed by point-to-point wireless links. We consider the SINR-based link transmission rate model involved in the slot-based link scheduling. We first prove a tight upper bound for the broadcast capacity of the multi-hop wireless networks under study. We then propose an online max-weight Throughput-Optimal Dynamic Broadcast algorithm (TODB) which performs link weight allocation, link scheduling, and data forwarding for each time slot according to the current network connectivity and data reception rates of nodes. We derive the throughput-optimality of the TODB algorithm by proving that the broadcast throughput of TODB can achieve the aforementioned upper bound of the broadcast capacity. We evaluate the performance of TODB in terms of throughput and latency via simulations and the results validate its effectiveness.

This article proposes a hybrid adaptive channel access (HyACA) mechanism for improving the channel access performance of an LTE-U and WiFi coexistence system. To support the HyACA mechanism, a modified time-frequency channel structure as well as a modified LBT functionality is first introduced for channel access in the coexistence system. To ensure channel access fairness, a dynamic channel switch (DCS) functionality and an adaptive almost-blank-subframe (AABS) functionality are further proposed to control the channel access under different traffic loads. Under low traffic load, HyACA uses the DCS functionality. Under high traffic load, HyACA uses the AABS functionality. To judge the traffic load status of the coexistence system, a criterion based on the average channel occupation time is introduced, and a calculation model is derived to determine the average channel occupation time. Moreover, another calculation model is derived to calculate the number of ABSs to be reserved in each LTE-U frame of a frame block. Through simulation results, it is demonstrated that the proposed HyACA mechanism can achieve better fairness and throughput performance than the static ABS mechanism, and better throughput performance than the LAA Cat.3 LBT mechanism while ensuring access fairness.

Mobile video traffic has experienced explosive growth in recent years due to the rapid development of mobile intelligent terminals and cellular communication technologies. The rapid growth of mobile video traffic has brought significant energy expenditure for mobile network operators. To reduce the energy expenditure, one promising solution is to exploit renewable energy harvested from surrounding environments for cellular traffic delivery. In this article, we investigate mobile video streaming in green cellular networks with hybrid energy sources, i.e., grid energy and ambient energy, to optimize both video quality and energy expenditure. Specifically, we formulate a stochastic optimization problem to maximize the long-term time-averaged network service utility, which is the difference of video quality and energy expenditure. The problem formulation takes the following factors into account: time-varying grid electricity price, energy harvesting process, and different time scales of rate adaptation (RA), resource management, and electricity price fluctuation. We exploit Lyapunov optimization framework to decompose the problem into three subproblems: 1) RA subproblem; 2) battery energy management subproblem; and 3) joint power control and subchannel assignment subproblem. We propose an efficient online green video streaming algorithm to solve these subproblems. We analyze the stability of the proposed algorithm with respect to lengths of energy queue and user request queues. Extensive simulations are conducted and the results validate the efficiency of the proposed algorithm.

In this paper, we study the design of data correlation aware opportunistic routing protocol for wireless sensor networks (WSNs) constituent of lossy links for improved network performance. We focus on a scenario where only M < N sensor nodes are required to report their readings, where N is the number of sensor nodes in the network. We formulate the problem of selecting M such reporting sensors among the N nodes as a mutual information maximization problem such that minimal information loss is resulted. Due to the NP-hardness of this issue, we present a greedy informative sensor selection algorithm. To reduce the transmission redundancy and improve the network lifetime performance, we design an energy efficient correlation aware hop-by-hop opportunistic routing protocol EECAR, which takes the following factors into account: spatial correlation characteristics among sensor readings, broadcast nature of wireless channels, and residual battery energy of sensor nodes. EECAR requires very limited routing information to be kept at nodes in the network. Extensive simulation results demonstrate the effectiveness of our greedy informative sensor selection algorithm and also show that EECAR outperforms existing work in terms of network lifetime and data gathering.