Delay-Aware and Energy-Efficient Computation Offloading in Mobile-Edge Computing Using Deep Reinforcement Learning

Ale, Laha; Zhang, Ning; Fang, Xiaojie; Chen, Xianfu; Wu, Shaohua; Li, Longzhuang

doi:10.1109/tccn.2021.3066619

Cited by 159 publications

(53 citation statements)

References 27 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…This makes the aforementioned classical optimization techniques very limited as they are mainly modeled based on a network snapshot and they must be reformulated when the dynamics changes over time. Besides, most of them need a high number of iterations and may provide a local rather than global optimum [8]. In fact, they mainly use heuristics to yield feasible solutions, which makes them subject to two major limitations [25] [26] [27]:…”

Section: Related Workmentioning

confidence: 99%

Deep Reinforcement Learning for the Computation Offloading in MIMO-based Edge Computing

Sadiki¹,

Dssouli²,

En‐Nouaary³

2021

Preprint

View full text Add to dashboard Cite

Multi-access Edge Computing (MEC) has recently emerged as a potential technology to serve the needs of mobile devices (MDs) in 5G and 6G cellular networks. By offloading tasks to high-performance servers installed at the edge of the wireless networks, resource-limited MDs can cope with the proliferation of the recent computationally-intensive applications. In this paper, we study the computation offloading problem in a massive multiple-input multiple-output (MIMO)-based MEC system where the base stations are equipped with a large number of antennas. Our objective is to minimize the power consumption and offloading delay at the MDs under the stochastic system environment. To this end, we formulate the problem as a Markov Decision Process (MDP) and propose two Deep Reinforcement Learning (DRL) strategies to learn the optimal offloading policy without any prior knowledge of the environment dynamics. First, a Deep Q-Network (DQN) strategy to solve the curse of the state space explosion is analyzed. Then, a more general Proximal Policy Optimization (PPO) strategy to solve the problem of discrete action space is introduced. Simulation results show that the proposed DRL-based strategies outperform the baseline and state-of-the-art algorithms. Moreover, our PPO algorithm exhibits stable performance and efficient offloading results compared to the benchmark DQN strategy.

show abstract

Section: Related Workmentioning

confidence: 99%

Deep Reinforcement Learning for the Computation Offloading in MIMO-based Edge Computing

Sadiki¹,

Dssouli²,

En‐Nouaary³

2021

Preprint

View full text Add to dashboard Cite

show abstract

“…An online algorithm based on Lyapunov optimization is proposed to jointly determine edge server site selection and energy harvesting. Work [16] investigated computation offloading in a dynamic MEC system with multiple edge servers, where computational tasks with various requirements were dynamically generated by IoT devices and then offloaded to MEC servers in a time-varying operating environment. e objective of this work is to maximize the task completion time and minimize the energy consumption of IoT devices.…”

Section: Related Workmentioning

confidence: 99%

“…According to the individual evaluation algorithm, all the individuals of rank 0 and rank 1 in the set P t B are put into the set P t 0 and P t 1 , respectively (8) If (P t 0 ≠ Φ) (9) select the optimal solution in P t 0 as the result and the algorithm ends (10) else (11) For P t B , the immune clonal selection algorithm is implemented to get P t I (12) Clear set P t 0 and P t 1 . According to the individual evaluation algorithm, all the individuals of rank 0 and rank 1 in the set P t B are put into the set P t 0 and P t 1 , respectively (13) If (P t 0 ≠ Φ) (14) select the optimal solution in P t 0 as the result and the algorithm ends (15) else (16) According to the individual evaluation algorithm, put the first n individuals in the set P t B ∪ P t I into P t+1 (17) t � t + 1 (18) end if (19) end if (20) end while (21) end if (22) select a solution randomly from P t 1 as the result and the algorithm ends ALGORITHM 2: Hybrid immune and bat scheduling algorithm (HIBSA). [33][34][35][36] 2 e upper frequency of bat fr min [33][34][35][36] 0 e lower frequency of bat α [37][38][39] 0.9 e update parameter of the loudness of bat c [37][38][39] 0. time of HIBSA is only 0.93. at is, the scheduling time of the algorithm proposed in this paper is only 93% of the TPGSA algorithm.…”

Section: Execution Time Analysismentioning

confidence: 99%

Computation Offloading Optimization in Mobile Edge Computing Based on HIBSA

Liu

Zhu

Wang

2021

Mobile Information Systems

View full text Add to dashboard Cite

Multiaccess edge computation (MEC) is a hotspot in 5G network. The problem of task offloading is one of the core problems in MEC. In this paper, a novel computation offloading model which partitions tasks into subtasksis proposed. This model takes communication and computing resources, energy consumption of intelligent mobile devices, and weight of tasks into account. We then transform the model into a multiobjective optimization problem based on Pareto that balances the task weight and time efficiency of the offloaded tasks. In addition, an algorithm based on hybrid immune and bat scheduling algorithm (HIBSA) is further designed to tackle the proposed multiobjective optimization problem. The experimental results show that HIBSA can meet the requirements of both the task execution deadline and the weight of the offloaded tasks.

show abstract

“…Motivated by these facts, we propose a multi-MEC server and multi-IoT device cellular network structure and investigate a weighted sum of multiple objectives minimization optimization problem in this paper. The weighted sum of multiple objectives optimization problems in dynamic MEC systems were also studied in [19][20][21], but the optimization objectives and system models are different to ours. The key differences between the relevant works and our work are shown in Table 1.…”

Section: Introductionmentioning

confidence: 99%

Edge Intelligence Empowered Dynamic Offloading and Resource Management of MEC for Smart City Internet of Things

et al. 2022

View full text Add to dashboard Cite

Internet of Things (IoT) has emerged as an enabling platform for smart cities. In this paper, the IoT devices’ offloading decisions, CPU frequencies and transmit powers joint optimization problem is investigated for a multi-mobile edge computing (MEC) server and multi-IoT device cellular network. An optimization problem is formulated to minimize the weighted sum of the computing pressure on the primary MEC server (PMS), the sum of energy consumption of the network, and the task dropping cost. The formulated problem is a mixed integer nonlinear program (MINLP) problem, which is difficult to solve since it contains strongly coupled constraints and discrete integer variables. Taking the dynamic of the environment into account, a deep reinforcement learning (DRL)-based optimization algorithm is developed to solve the nonconvex problem. The simulation results demonstrate the correctness and the effectiveness of the proposed algorithm.

show abstract

Delay-Aware and Energy-Efficient Computation Offloading in Mobile-Edge Computing Using Deep Reinforcement Learning

Cited by 159 publications

References 27 publications

Deep Reinforcement Learning for the Computation Offloading in MIMO-based Edge Computing

Deep Reinforcement Learning for the Computation Offloading in MIMO-based Edge Computing

Computation Offloading Optimization in Mobile Edge Computing Based on HIBSA

Edge Intelligence Empowered Dynamic Offloading and Resource Management of MEC for Smart City Internet of Things

Contact Info

Product

Resources

About