关键词： Autonomous aerial vehicles   Optimization   Mobile handsets   Servers   Delays   Relays   Heuristic algorithms   Multi-access edge computing   Computer architecture   Cloud computing   Block coordinate descent   computation offloading   mobile edge computing   unmanned aerial vehicle deployment  

摘要： As a promising technique for offloading computation tasks from mobile devices, Unmanned Aerial Vehicle (UAV)-assisted Mobile Edge Computing (MEC) utilizes UAVs as computational resources. A popular method for enhancing the quality of service (QoS) of UAV-assisted MEC systems is to jointly optimize UAV deployment and computation task offloading. This imposes the challenge of dynamically adjusting UAV deployment and computation offloading to accommodate the changing positions and computational requirements of mobile devices. Due to the real-time requirements of MEC computation tasks, finding an efficient joint optimization approach is imperative. This paper proposes an algorithm aimed at minimizing the average response delay in a UAV-assisted MEC system. The approach revolves around the joint optimization of UAV deployment and computation offloading through convex optimization. We break down the problem into three sub-problems: UAV deployment, Ground Device (GD) access, and computation tasks offloading, which we address using the block coordinate descent algorithm. Observing the $NP$NP-hardness nature of the original problem, we present near-optimal solutions to the decomposed sub-problems. Simulation results demonstrate that our approach can generate a joint optimization solution within seconds and diminish the average response delay compared to state-of-the-art algorithms and other advanced algorithms, with improvements ranging from 4.70% to 42.94%.

关键词： Delays   Multitasking   Internet of Things   Artificial neural networks   Task analysis   Computational modeling   Inference algorithms   Asynchronous advantage actor-critic   distributed DNN inference   mobile edge computing   model partitioning   multi-task learning  

摘要： Model partitioning is a promising technique for improving the efficiency of distributed inference by executing partial deep neural network (DNN) models on edge servers (ESs) or Internet-of-Things (IoT) devices. However, due to heterogeneous resources of ESs and IoT devices in mobile edge computing (MEC) networks, it is non-trivial to guarantee the DNN inference speed to satisfy specific delay constraints. Meanwhile, many existing DNN models have a deep and complex architecture with numerous DNN blocks, which leads to a huge search space for fine-grained model partitioning. To address these challenges, we investigate distributed DNN inference with fine-grained model partitioning, with collaborations between ESs and IoT devices. We formulate the problem and propose a multi-task learning based asynchronous advantage actor-critic approach to find a competitive model partitioning policy that reduces DNN inference delay. Specifically, we combine the shared layers of actor-network and critic-network via soft parameter sharing, and expand the output layer into multiple branches to determine the model partitioning policy for each DNN block individually. Experiment results demonstrate that the proposed approach outperforms state-of-the-art approaches by reducing total inference delay, edge inference delay and local inference delay by an average of 4.76%, 10.04% and 8.03% in the considered MEC networks.

关键词： Computational modeling   Training   Data models   Costs   Privacy   Mobile computing   Degradation   Differential privacy   federated learning   lottery ticket hypothesis   mobile edge computing   zero -concentrated DP  

摘要： Federated learning (FL) can train collaboratively with several mobile terminals (MTs), which faces critical challenges in communication, resource, and privacy. Existing privacy-preserving methods usually adopt instance-level differential privacy (DP), which provides a rigorous privacy guarantee but with several bottlenecks: performance degradation, transmission overhead, and resource constraints. Therefore, we propose Fed-LTP, an efficient and privacy-enhanced FL framework with Lottery Ticket Hypothesis (LTH) and zero-concentrated DP (zCDP). It generates a pruned global model on the server side and conducts sparse-to-sparse training from scratch with zCDP on the client side. On the server side, two pruning schemes are proposed: (i) the weight-based pruning (LTH) determines the pruned global model structure;(ii) the iterative pruning further shrinks the size of the pruned model. Meanwhile, the performance of Fed-LTP is boosted via model validation based on the Laplace mechanism. On the client side, we use sparse-to-sparse training to solve the resource-constraints issue and provide tighter privacy analysis to reduce the privacy budget. We evaluate the effectiveness of Fed-LTP on several real-world datasets in both independent and identically distributed (IID) and non-IID settings. The results confirm the superiority of Fed-LTP over state-of-the-art (SOTA) methods in communication, computation, and memory efficiencies while realizing a better utility-privacy trade-off.

关键词： Task analysis   Vehicle dynamics   Resource management   Delays   Energy consumption   Costs   Dynamic scheduling   Deep reinforcement learning   mobile edge computing   resource allocation   task offloading  

摘要： The emergence of mobile edge computing (MEC) technology opens up new opportunities for the realization of demanding applications proximate to the user. This work aims to capitalize on the constrained resources of MEC networks to efficiently fulfill service requests by jointly optimizing task offloading and resource allocation. However, the growing complexity of applications such as intelligent transportation and autonomous driving introduces new challenges in modeling and problem-solving due to the causal correlation of service logic among various users, which adds to the intricate spatial-temporal dynamics of the network. To effectively capture the essence of cross-user task dependency in these scenarios while also accommodating dynamic factors like the variability of wireless channels and the mobility of terminals, we formulate a spatial-temporal correlative problem, aiming to minimize the weighted cost of task completion delay and energy consumption. To resolve this problem, a Double Deep Q Network (DDQN) is developed for task offloading while integrating subchannel allocation and transmit power control into it to interact with the dynamic environment to generate a reward signal, thereby optimizing the long-term performance of the system. Comprehensive simulations verify that the proposed algorithm surpasses the comparative methods in reducing the cost. Additionally, adaptability in different scenarios with different parameters and task structures has also been validated and analyzed.

关键词： 5G   Mobile edge computing   Queue system   Stochastic geometry  

摘要： With the development of communication and computation technologies, mobile edge computing (MEC) has become an emerging technology being deployed in 5G/B5G mobile networks to improve the mobile user experience. In this paper, we study the asymptotic performance of parallel computing in MEC networks with massive multi -input and multi -output (MIMO) in the millimeter wave band. First, a massive MIMO enabled MEC network with a parallel -limited computing model is constructed, and the task completed probability is defined as a new network performance metric. In the proposed model, we derive the task offloading and computing probabilities via the probability generation functional of the Poisson point process, from which, the task completed probability is obtained. Then we propose a computing resource allocation optimization algorithm to maximize the task completed probability within a given delay constraint. The numerical simulation results prove the accuracy of the obtained theoretical results and verify the performance of the proposed algorithm.

关键词： Mobile handsets   Task analysis   Computational modeling   Inference algorithms   Pipelines   Partitioning algorithms   Multi-access edge computing   Collaborative inference   deep reinforcement learning   mobile edge computing   model partition   service migration  

摘要： Deep neural networks are the cornerstone of many mobile intelligent systems, and their inference processes bring about computation-intensive tasks. Device-edge cooperative inference in mobile edge computing provides a fine-grained processing method to migrate the burden of inference computation. However, the geographical dispersion of resources and the mobility pattern of devices pose scheduling issues to be considered. In this paper, we propose a task scheduling framework for such device-edge systems to improve the pipeline time of model inference. First, we consider the resource provisioning strategy with a pre-fetching service migration setting in the environment of multiple mobile devices and edge nodes. Then, we leverage game theory to analyze the property of the decision-making process and propose an offline algorithm under complete information. Next, we propose an algorithm based on proximal policy optimization to enable mobile devices to make decisions in a distributed online manner. Further, we adopt a memory mechanism into the online algorithm to improve the decision-makers' understanding of the system environment. Experiments demonstrate the effectiveness of the two algorithms. The average pipeline time of the proposed online algorithm is only 61.44% of that of local processing, which is 1.196 times that of the proposed offline algorithm.

关键词： Mobile edge computing   Computation scheduling   Energy harvesting   Task offloading  

摘要： The rapid expansion of Mobile Edge Computing (MEC) is driven by the growing demand for resource-intensive applications within the Internet of Things. Computation offloading allows these applications to be executed at the network edge, but it leads to significant electricity expenses for network operators. To mitigate these costs, one promising solution is to power base stations (BSs) with a hybrid energy supply that combines unpredictable harvested energy with stable energy from the smart grid. This paper investigates joint computation scheduling for mobile devices (MDs) and resource allocation in a MEC network incorporating hybrid energy sources. Our objective is to maximize long-term time-averaged service utility by optimizing parameters such as BS battery supply, harvestable energy, CPU frequency, transmission power, task-partition factor, and MD-BS associations. To tackle this complex problem, we exploit the Lyapunov optimization framework to decompose it into deterministic subproblems for each time slot and propose an online network service utility maximization scheduling (NSUMS) algorithm. Experimental results show that our algorithm outperforms benchmark schemes in service utility and energy expenditure, improving the completion ratio by 32%, reducing the failure rate by 80%, and decreasing MD energy consumption by 28%.

关键词： Task analysis   Heuristic algorithms   System dynamics   Measurement   Data processing   Minimization   Servers   Age of information   mobile edge computing   post-decision state   deep reinforcement learning  

摘要： With the rapid development of Mobile Edge Computing (MEC), various real-time applications have been deployed to benefit people's daily lives. The performance of these applications relies heavily on the freshness of collected environmental information, which can be quantified by its Age of Information (AoI). In the traditional definition of AoI, it is assumed that the status information can be actively sampled and directly used. However, for many MEC-enabled applications, the desired status information is updated in an event-driven manner and necessitates data processing. To better serve these applications, we propose a new definition of AoI and, based on the redefined AoI, we formulate an online AoI minimization problem for MEC systems. Notably, the problem can be interpreted as a Markov Decision Process (MDP), thus enabling its solution through Reinforcement Learning (RL) algorithms. Nevertheless, the traditional RL algorithms are designed for MDPs with completely unknown system dynamics and hence usually suffer long convergence times. To accelerate the learning process, we introduce Post-Decision States (PDSs) to exploit the partial knowledge of the system's dynamics. We also combine PDSs with deep RL to further improve the algorithm's applicability, scalability, and robustness. Numerical results demonstrate that our algorithm outperforms the benchmarks under various scenarios.

关键词： Mobile edge computing   Multi-objective optimization   Evolutionary algorithms   Energy consumption   Delay  

摘要： In mobile edge computing (MEC), edge clients have the capability to offload their computing tasks to edge servers or other edge clients, aiming to reduce their energy consumption and latency. However, the process of offloading introduces additional energy consumption and time due to long-distance network transmission, which presents a significant challenge in jointly minimizing energy consumption and latency for edge clients. To address this challenge, we formulate the problem as a constrained multi-objective optimization problem and propose an enhanced multi-objective evolutionary algorithm, named MOEA-TOMEC, to determine optimal task offloading strategies in MEC. The proposed algorithm aims to minimize both the extra energy consumption and latency. To overcome the limitations of a single offloading mode in complex network environments, we introduce a hybrid offloading mode within our algorithm. This hybrid approach combines partial offloading and full offloading to achieve lower energy consumption and latency. Additionally, we propose a device classification algorithm that plays a crucial role in initializing the population. This algorithm incorporates low-battery devices as relays in the offloading decisions, thereby enhancing population diversity. By generating a superior initial population and expanding the breadth of the learning space, our proposed algorithm achieves improved performance. Experimental results demonstrate the effectiveness of our proposed algorithm. In comparison to existing algorithms, our approach reduces the energy consumption and latency of edge clients by an average of 9.13% and 13.91%, respectively.

关键词： Mobile edge computing   Cloud server   5G networks   Non-orthogonal multiple access   Telecommunication resources  

摘要： Mobile edge processing is a cutting-edge technique that addresses the limitations of mobile devices by enabling users to offload computational tasks to edge servers, rather than relying on distant cloud servers. This approach significantly reduces the latency associated with cloud processing, thereby enhancing the quality of service. In this paper, we propose a system in which a cellular network, comprising multiple users, interacts with both cloud and edge servers to process service requests. The system assumes non-orthogonal multiple access (NOMA) for user access to the radio spectrum. We model the interactions between users and servers using queuing theory, aiming to minimize the total energy consumption of users, service delivery time, and overall network operation costs. The problem is mathematically formulated as a multi-objective, bounded non-convex optimization problem. The Structural Correspondence Analysis (SCA) method is employed to obtain the global optimal solution. Simulation results demonstrate that the proposed model reduces energy consumption, delay, and network costs by approximately 50%, under the given assumptions.