29 October 2021 in IEEE Transactions on Network and Service Management
Edge computing is emerging to empower the future of Internet of Things (IoT) applications. However, due to heterogeneity of applications, it is a significant challenge for the edge cloud to effectively allocate multidimensional limited resources (CPU, memory, storage, bandwidth, etc.) with constraints of applications’ Quality of Service (QoS) requirements. In this paper, we address the resource allocation problem in Edge-IoT systems through developing a novel framework named DeepEdge that allocates resources to the heterogeneous IoT applications with the goal of maximizing users’ Quality of Experience (QoE). To achieve this goal, we develop a novel QoE model that considers aligning the heterogeneous requirements of IoT applications to the available edge resources. The alignment is achieved through selection of QoS requirement range that can be satisfied by the available resources. In addition, we propose a novel two-stage deep reinforcement learning (DRL) scheme that effectively allocates edge resources to serve the IoT applications and maximize the users’ QoE. Unlike the typical DRL, our scheme exploits deep neural networks (DNN) to improve actions’ exploration by using DNN to map the Edge-IoT state to joint resource allocation action that consists of resource allocation and QoS class. The joint action not only maximize users’ QoE and satisfies heterogeneous applications’ requirements but also align the QoS requirements to the available resources. In addition, we develop a Q-value approximation approach to tackle the large space problem of Edge-IoT. Further evaluation shows that DeepEdge brings considerable improvements in terms of QoE, latency and application tasks’ success ratio in comparison to the existing resource allocation schemes.
23 January 2018 in IEEE Transactions on Mobile Computing
5G is the upcoming evolution for the current cellular networks that aims at satisfying the future demand for data services. Heterogeneous cloud radio access networks (H-CRANs) are envisioned as a new trend of 5G that exploits the advantages of heterogeneous and cloud radio access networks to enhance spectral and energy efficiency. Remote radio heads (RRHs) are small cells utilized to provide high data rates for users with high quality of service (QoS) requirements, while high power macro base station (BS) is deployed for coverage maintenance and low QoS users service. Inter-tier interference between macro BSs and RRHs and energy efficiency are critical challenges that accompany resource allocation in H-CRANs. Therefore, we propose an efficient resource allocation scheme using online learning, which mitigates interference and maximizes energy efficiency while maintaining QoS requirements for all users. The resource allocation includes resource blocks (RBs) and power. The proposed scheme is implemented using two approaches: centralized, where the resource allocation is processed at a controller integrated with the baseband processing unit and decentralized, where macro BSs cooperate to achieve optimal resource allocation strategy. To foster the performance of such sophisticated scheme with a model free learning, we consider users' priority in RB allocation and compact state representation learning methodology to improve the speed of convergence and account for the curse of dimensionality during the learning process. The proposed scheme including both approaches is implemented using software defined radios testbed. The obtained results and simulation results confirm that the proposed resource allocation solution in H-CRANs increases the energy efficiency significantly and maintains users' QoS.
25 July 2017 in IEEE Transactions on Vehicular Technology
The multitier heterogeneous structure of 5G with dense small cells deployment, relays, and device-to-device (D2D) communications operating in an underlay fashion is envisioned as a potential solution to satisfy the future demand for cellular services. However, efficient power allocation among dense secondary transmitters that maintains quality of service (QoS) for macro (primary) cell users and secondary cell users is a critical challenge for operating such radio. In this paper, we focus on the power allocation problem in the multitier 5G network structure using a noncooperative methodology with energy efficiency consideration. Therefore, we propose a distributive intuition-based online learning scheme for power allocation in the downlink of the 5G systems, where each transmitter surmises other transmitters power allocation strategies without information exchange. The proposed learning model exploits a brief state representation to account for the problem of dimensionality in online learning and expedite the convergence. The convergence of the proposed scheme is proved and numerical results demonstrate its capability to achieve fast convergence with QoS guarantee and significant improvement in system energy efficiency.
14 May 2019 in IEEE Transactions on Green Communications and Networking
The energy efficiency in multi-tier heterogeneous 5G networks is a critical issue due to the fact that the macro base stations (BSs) power consumption is considerably high and proportional to their traffic load. Traffic offloading from macrocells to small cells is envisioned as a potential solution to improve energy efficiency in 5G heterogeneous networks. However, traffic offloading causes traffic congestion at small cells, interference as a result of small cells transmissions, and aggregate power consumption of small cells. These factors make the traffic offloading procedure more challenging. In this paper, we propose a conditional traffic offloading scheme, which relies on macrocells and small cells system load information to determine the most energy efficient traffic offloading strategy, select the proper operation mode of small cells, and fulfill macro users applications’ quality of service (QoS) requirements. The proposed scheme is developed using a novel intuitive online reinforcement learning methodology to perform the conditional traffic offloading in which each macro BS conjectures the offloading strategies of other macrocells. The convergence of the proposed scheme is proved and numerical results demonstrate its capability to achieve fast convergence with QoS guarantee and significant energy efficiency.