By integrating multi-dimensional capability elements such as communication, sensing, computing and AI, the fusion of telepathy and computing can enable multi-functional collaboration, resource sharing, cost reduction, and stronger synergies. The traditional service caching and task offloading strategies for mobile edge computing (MEC) no longer adapt to the competition and coupling between the performance matching of multi-dimensional resources in the MEC network oriented to telepathy and computing. To this end, the service caching and task offloading model of mobile edge computing fused with telepathy and computing is studied and established; considering the time-varying network environment and incomplete observation state, under the constraints of multi-dimensional resources such as perception accuracy, computing power, storage and energy consumption, the joint optimization problem of service caching, power control and task offloading is abstracted into a partially observable Markov decision process with the goal of minimizing task computing delay and energy consumption; while ensuring the perception accuracy, a service caching and task offloading strategy based on cyclic multi-agent deep reinforcement learning is proposed. The simulation results show that the algorithm has significant performance improvements in terms of service cache hit rate.

Uncrewed aerial vehicles (UAVs), due to their inherent flexibility and autonomous operation, are widely used in Internet of Things (IoT) networks for collecting data to facilitate real-time evaluation and monitoring applications. This article investigates a UAV-assisted IoT network, in which the UAV sequentially accesses IoT devices (IoTDs). During hovering, the UAV works on a full-duplex mode while collecting data from target devices, taking into account actual propulsion power consumption. In order to quantify the freshness of devices data, we introduce the concept of Age of Information (AoI). A multiobjective optimization method is proposed to jointly optimize three objectives: 1) maximization of the data rate; 2) minization of AoI; and 3) minization of the UAV energy consumption over a particular mission period. These three objectives partially conflict with each other and provide weight parameters to describe their importance. Since the data uploaded by IoTDs is dynamically changing, the trajectory planning of the UAV is required. Considering that the UAV has no prior knowledge of the network environment, the optimization problem is reformulated as a Markov decision process. Aiming at the learning problem of the UAV control strategies over multiple objectives, a deep reinforcement learning algorithm for multiobjective collaborative optimization is proposed. While training, the agent collects data in time according to the devices priority, and generates optimal strategies under the conditions of giving weights. Experimental results demonstrate that the proposed multiobjective twin delayed deep deterministic policy gradient algorithm jointly optimizes three objectives and can adjust the optimal policy based on the weight parameters of each objective.

With the rapid advancement of Internet of Things (IoT), there is a growing demand for intelligent applications with varying requirements (e.g. delay, reliability, and energy). Mobile Edge Computing (MEC) can enhance the responsiveness of these applications by caching specific computing services on edge servers. However, the efficiency of task offloading and service caching at MEC servers is often hindered by the diversity of user preferences and privacy across different edge node service regions, especially in heterogeneous networks. To this end, we introduce a personalized computation offloading and service caching method that integrates Deep Reinforcement Learning (DRL) with Personalized Federated Learning (PFL), termed DPFL, aimed at optimizing computation performances in a distributed and privacy-preserving manner. The DPFL employs DRL to jointly optimize computation offloading and service caching placement, reducing task latency and energy consumption. Simultaneously, it leverages personalized federated learning to develop local service prediction models, offering tailored service caching policies for users in heterogeneous regions while safeguarding data privacy. Simulation results demonstrate that our algorithm surpasses existing policies in reducing application delay, energy consumption, and improving system cache hit ratio.

In response to the high computational performance demands of emerging compute-intensive applications for uncertain internet of things (IoT) scenarios, this paper proposes a task offloading and resource allocation algorithm combined with deep reinforcement learning. Firstly, under the constraints of computing resources of mobile devices (MDs) as well as IoT terminals and latency computing tasks, a mixed-integer nonlinear programming (MINLP) problem for joint task offloading and resource allocation is established. Due to the time varying nature of dynamic MEC scenarios and the fact that single agent could only access partial envirionmental state, the original problem is transformed into a partially observable markov decision process (POMDP). However, in dense and uncertain scenarios, each MD has distinct requirements in terms of latency and the computational capacity of Edge Servers (ES) during offloading. To this end, a reliability-enhanced MADDPG (RE-MADDPG) algorithm is proposed to generate the user offloading strategy according to the delay and reliability without knowing the information of the MEC server in the current slot. The simulation results show that the proposed algorithm can effectively reduce the delay and provide users with more efficient services.

The multi-access edge computing (MEC) and ultra-dense network (UDN) are regarded as essential and complementary technologies in the age of Internet of Things (IoT). Deploying MEC servers at the macro-cell and small-cell stations can significantly improve user experience as well as increase network capacity. Nevertheless, there still remain many obstacles in practical MEC-enabled UDNs. Among them, a unique challenge is how to coordinate computing and networking to fit the diverse offloading demands of IoT applications in dynamic network environments. To this end, this paper first investigates a distributed delay-constrained computation offloading methodology based on computing and networking coordination in the UDN. An extended game-theoretic approach based on the Lyapunov optimization theory is designed to achieve adaptive task offloading and computing power management in time-varying environments. Furthermore, considering the uncertainty in users' mobility and limited edge resources, distributed two-stage and multi-stage stochastic programming algorithms under various uncertainties are proposed. The proposed algorithms take posterior recourse actions to compensate for inaccurate predicted network information. Extensive simulations validate the effectiveness and rationality of the proposed algorithms and their superior performance over several benchmark schemes.

Mobile edge computing (MEC) enables various services to be cached in close proximity to the user equipments (UEs), thereby reducing the service delay of many emerging applications. However, the limitation of storage, computation, and radio resources, the dynamics of the decentralized MEC environment, and the complex spatial relationships of service request types and wireless network states between edge nodes make it difficult to realize efficient edge computing services. To address these challenges, this paper integrates the digital twin (DT) technology with a multi-cell MEC network to study an intelligent cooperative task offloading and service caching scheme, aiming at maximizing a quality of services (QoE)-based system utility. Specifically, we first construct a digital twin edge network (DITEN) to reflect the physical MEC system in real-time and provide data for training. With the help of DT technology, it is easy to access data resources in the DITEN to improve the simulation ability and reduce the communication cost. Then, we propose a graph attention-based multi-agent reinforcement learning (GatMARL) algorithm to learn the optimal task offloading and service caching strategies in the DITEN. The GatMARL employs a graph attention-based value decomposition network to capture the potential spatial relationships between edge nodes to learn better attentive cooperation policy. Simulation results demonstrate that the proposed GatMARL algorithm exhibits an effective performance improvement compared with state-of-the-art benchmarks.

Accurate spatial-temporal traffic modeling and prediction play an important role in intelligent transportation systems (ITS). Recently, various deep learning methods such as graph convolutional networks (GCNs) and recurrent neural networks (RNNs) have been widely adopted in traffic prediction tasks to extract spatial-temporal dependencies based on a large volume of high-quality training data. However, there exist data scarcity problems in some transportation networks, and in these cases, the performance of traditional GCNs and RNNs based approaches will degrade sharply. To address this problem, this paper proposes an adversarial domain adaptation with spatial-temporal graph convolutional network (Ada-STGCN) model to predict traffic indicators for a data-scarce target road network by transferring the knowledge from a data-sufficient source road network. Specifically, Ada-STGCN first develops a spatial-temporal graph convolutional network that combines the GCN and gated recurrent unit (GRU) to extract spatial-temporal dependencies from source and target road networks. Then, the technique of adversarial domain adaptation is integrated with the spatial-temporal graph convolutional network to learn discriminative and domain-invariant features to facilitate knowledge transfer. Experimental results on the real-world traffic datasets in the traffic flow prediction task demonstrate that our model yields the best prediction performance compared to state-of-the-art baseline methods.

Mobile edge computing (MEC) enables various services to be cached in close proximity to the user equipments (UEs), thereby reducing the computing delay of many emerging applications. Nevertheless, The limited storage capacity of edge servers requires judicious design of service caching as well as task offloading to maximize edge computing performances. In this paper, we formulate a cooperative task offloading, service caching, and transmit power allocation problem to minimize the cost of computing delay and energy consumption of UEs. To address this problem, we propose a graph attention based multi-agent deep deterministic policy gradient (GAT-MADDPG) algorithm, in which a multi-headed graph attention mechanism is incorporated into the centralized critic network to learn the attentive cooperation policies. Simulation results show that the proposed GAT-MADDPG algorithm exhibits an effective performance improvement.

In order to improve mobile users’ service experience, mobile cloud computing (MCC) is promoted. Although MCC can alleviate the burdens of Smart mobile devices (SMDs) by offloading computation-intensive applications to the cloud, it also aggravates computing and storage overheads in cloud centers and bandwidth overhead on wireless links for offloading workloads of mobile users. Therefore, we should carefully design the offloading policy to decrease these overheads while easing the burdens of SMDs. To this end, we investigate the offloading policy in heterogeneous wireless networks. In this paper, a queue model is built to formulate the mobile users’ workload offloading problem and Lyapunov optimization framework is proposed to make trade-off between system offloading utility and queue backlog. For deterministic WiFi connections, a Lagrangian optimization method is proposed to decide the optimal offloading workloads. Furthermore, considering random WiFi connection durations, a multi-stage stochastic programming method is proposed. The experimental results show effectiveness of the Lagrangian optimization offloading method for deterministic WiFi connection and the multi-stage stochastic programming method for random WiFi connection.

With the rapid advancement of the Internet of Vehicles and artificial intelligence (AI) technologies, the cooperative intelligent transportation system (C-ITS) has drawn great attention in recent years. To provide an ultra-reliable, low-latency computation experience of C-ITS, computation offloading is deemed indispensable by working with edge-cloud servers. In this paper, we first investigate a distributed dynamic computation offloading model for multi-access edge computing (MEC) enabled C-ITS under a heterogeneous road network, in which the multiple and heterogeneous computing power sources cooperatively provide computation offloading services for vehicles. Considering the autonomous offloading manner of the vehicles, we formulate the task offloading and computing power allocation as a distributed Stackelberg game, where the MEC servers as the leader to allocate computing resources and manage local energy, and the vehicles as the followers to offload local computation task. Since the observable states in the game is incomplete, the problem of resolving the optimal strategies for each game player is modeled as a partially observable Markov decision process (POMDP) to maximize the long-term cumulative reward. Then we develop a computation offloading algorithm using Stackelberg game-based multi-agent deep deterministic policy gradient (SG-MADDPG), which uses a centralized training and decentralized execution method to learn the optimal computing power allocation and computation offloading policies. Finally, extensive simulations are carried out and show the rationality and effectiveness of the proposed algorithm.

Resources allocation efficiently of wireless network virtualization (WNV) in Internet of Things (IoT) has become a key challenge owing to the conflict between the limited capacity of physical resources and the massive virtual resources requests. Focusing on time-frequency resource allocation in WNV-based IoT, this article proposes a dynamic resource allocation algorithm based on the lifetime priority (LP-DRA) of virtual network requests (VNRs). First, LP-DRA investigates the continuity of physical time-frequency resource blocks by the Karnaugh map approach, then calculates the reallocation impact factor (RIF) of existing virtual networks (VNs), and reallocates the time-frequency resources to the VN with the largest RIF. Finally, the resources are preferentially allocated to the VNs with shorter lifetimes. The performance of LP-DRA is compared with the static resource allocation algorithm and the greedy dynamic resource allocation algorithm. The simulation results show that LP-DRA can utilize physical resources more effectively, improve the acceptance rate of VNRs, and increase the revenue of the physical network.

With the rapid development and convergence of the mobile Internet and the Internet of Things (IoT), computing-intensive and delay-sensitive IoT applications (APPs) are proliferating with an unprecedented speed in recent years. Mobile edge computing (MEC) and energy harvesting (EH) technologies can significantly improve the user experience by offloading computation tasks to edge-cloud servers as well as achieving green and durable operation. Traditional centralized strategies require precise information of system states, which may not be feasible in the era of big data and artificial intelligence. To this end, how to allocate limited edge-cloud computing resource on demand, and how to develop heterogeneous task offloading strategies with EH in a more flexible manner are remaining challenges. In this paper, we investigate an EH-enabled MEC offloading system, and propose an online distributed optimization algorithm based on game theory and perturbed Lyapunov optimization theory. The proposed algorithm works online and jointly determines heterogeneous task offloading, on-demand computing resource allocation, and battery energy management. Furthermore, to reduce the unnecessary communication overhead and improve the processing efficiency, an offloading pre-screening criterion is designed by balancing battery energy level, latency, and revenue. Extensive simulations are carried out to validate the effectiveness and rationality of the proposed approach.

With the rapid development of Internet of Things (IoT) on e-Health, the role of Mobile Edge Computing (MEC) has been increasingly effective in providing high-performance, lowlatency computing services. In this work, we consider the problem of task offloading and computing resource allocation in dynamic environment, wherein heterogeneous IoT devices or e-Health applications with diverse requirements in latency and energy constraint. Taking into account the different traffic characteristics and spatio-temporally varying distributed environment, we formulate the offloading problem as a dynamic game and a Stackelberg Equilibrium (SE) based distributed online offloading manner is proposed. And then, to allocate computing resource on demand, a dynamic quote price mechanism is designed by invoking Lyapunov optimization. Furthermore, to improve processing efficiency and reduce unnecessary communication overhead, a “first-rank” servers selection criteria is proposed by balancing revenue and latency. Finally, the effectiveness and rationality of the algorithm are verified by experimental simulation.

With the exponentially increasing number of mobile devices, crowdsensing has been a hot topic to use the available resource of neighbor mobile devices to perform sensing tasks cooperatively. However, there still remain three main obstacles to be solved in the practical system. First, since mobile devices are selfish and rational, it is natural to provide cooperation for sensing with a reasonable payment. Meanwhile, due to the arrival and departure of sensing tasks, resource should be allocated and released dynamically when sensing task comes or leaves. To this end, this paper designs a game theoretic approach based incentive mechanism to encourage the “best” neighbor mobile devices to share their own resource for sensing. Next, in order to adjust resource among mobile devices for the better crowdsensing response, an auction based task migration algorithm is proposed, which can guarantee the truthfulness of announced price of auctioneer, individual rationality, profitability, and computational efficiency. Moreover, taking into account the random movement of mobile devices resulting in the stochastic connection, we also use multi-stage stochastic decision to take posterior resource allocation to compensate for inaccurate prediction. The numerical results show the effectiveness and improvement of the proposed multi-stage stochastic programming based distributed game theoretic methodology (SPG) for crowdsensing.

Offloading has been widely adopted as an effective technique to overcome the processing and computation limitation in mobile networks. In this work, we consider the problem of offloading in a mobile ad hoc environment in order to improve the processing capability and power efficiency. We formulate this as an incentive-based workload assignment problem. For maximizing the individual utility, the buyer/seller game is formulated to model the interactions among mobile devices for offloading. We derive the Stackelberg Equilibrium solution is to determine the workload assignment and power allocation. Based on this, we design distributed allocation algorithms, and the results from experiment verify the effectiveness of our proposal.