Relying on the computation offloading technology, edge computing has shown potential in countless tasks processing in the industrial Internet of Things (IIoT), which is composed of multiple edge clouds and multiple IIoT devices. Nevertheless, with increasing demands for computation service, how to design reliable transmission mechanism and allocate proper computation resource has become bottlenecks. In this paper, we propose a computation offloading mechanism based on two-stage Stackelberg game to analyze the interaction between multiple edge clouds and multiple IIoT devices. To be specific, the edge clouds are denoted as leaders who set the appropriate price for their computation resource. Besides considering the payment cost, the IIoT devices who are termed as the followers formulate their utility function by considering the social interaction information from the potential IIoT devices. The existence and uniqueness of the Stackelberg equilibrium are analyzed considering two possible cases, i.e. complete information and incomplete information. Moreover, two dynamic iterative algorithms are invoked for solving the both problem models, respectively. Finally, experimental results show that our proposed scheme is conducive to seeking the appropriate price and computation requirement. Besides, social interaction information plays an important role in achieving a reasonable computation requirement for IIoT devices.
24 September 2020 in IEEE Transactions on Mobile Computing
Relying on ofloading computation tasks to the network edge, ultra dense networks (UDNs) are capable of providing delay-aware service to nearby users. Meanwhile, software defined networking (SDN) is deemed as an effective technology to ease the management of infrastructure plane and control plane in UDNs. Specifically, the centralized SDN controller is capable of managing the whole network globally. With the increasing demands for various applications as well as the limitation of computation, storage and communication resource, how to allocate spectrum resource appropriately is imperative. In this paper, we mainly show solicitude for spectrum sharing and edge computation ofloading problems in SDN-based ultra dense networks. To address this issue, we propose a second-price auction scheme for ensuring the fair bidding for spectrum rent. Moreover, the MBS edge cloud is termed as the buyer, and the SBS edge clouds are the sellers who sell the ofloading resource to the MBS edge cloud. To be specific, the spectrum sharing and computation ofloading scheme is executed in the SDN controller, and the controller is responsible for distributing spectrum allocation instructions to the infrastructure plane. Finally, experimental results validate the effectiveness of our proposed scheme in SDN-based ultra dense networks.
Given decoupling the control layer and the infrastructure layer, the software defined wireless networks (SDWN) is beneficial in terms of providing both low latency and low energy consumption services for mobile users, where multicontroller placement and resource management become a pair of bottlenecks. In this paper, we propose an energy-aware multicontroller placement scheme as well as a latency-aware resource management model for the SDWN. Moreover, the particle swarm optimization (PSO) is invoked for solving the multi-controller placement problem, and a deep reinforcement learning (DRL) algorithm aided resource allocation strategy is conceived. Finally, experimental results show that our proposed schemes are conducive to reducing both the execution time and the energy consumption of each task.
1 January 2016 in International Journal of Wireless and Mobile Computing
Support vector machine (SVM) model is becoming an increasingly popular method in software defects prediction. This model has strong non-linear classifying ability. However, SVM model lacks effective method to determine the best parameters. In this paper, a modified bat algorithm, named changing range bat algorithm, is employed to optimise the parameters of SVM model. To test the performance of this new model, several public datasets of software defect prediction are employed and then the results are compared with other five approaches. Experimental results show that the classification ability of hybrid CRBA-SVM model surpasses all other approaches.
With the increasing dissemination of the Internet of Things and 5G, mobile edge computing has become a novel scheme to assist terminal devices in executing computation tasks. To elevate the coverage and computation capability of edge computing, a collaborative computation offloading and resource management architecture was proposed in space–air–ground integrated networking (SAGIN). In this manuscript, we established a novel model considering the computation offloading cost constraints of the communication, computing and cache model in the SAGIN. To be specific, the joint optimization problem of collaborative computation offloading and resource management was modeled as a mixed integer nonlinear programming problem. To address this issue, this paper proposed a computation offloading and resource allocation strategy based on deep reinforcement learning (DRL). Differing from traditional methods, DRL does not need a well-established formulation or previous information, and it is capable of revising the strategy adaptively according to the environment. The simulation results demonstrate the proposed approach can achieve the optimal reward values in the case of different terminal device numbers. Furthermore, this manuscript provided the analysis with variant parameters of the proposed approach.