The wide applications of artificial intelligence (AI) on massive Internet-of-things or smartphones raises significant concerns about privacy, heterogeneity, and resource efficiency. Correspondingly, federated learning emerges as an effective way to enable AI over massively distributed nodes without uploading the raw data. Conventional works mostly focus on learning a single unified model for one solitary task. Multi-task learning (MTL) outperforms single-task learning by training multiple models concurrently, leading to reduced model sizes and increased flexibility. However, existing federated learning efforts often face challenges in efficiently managing MTL scenarios, particularly with the presence of stragglers, without incurring prohibitive computation and communication costs. In this paper, inspired by the natural cloud-BS-terminal hierarchy of cellular networks, we provide a viable resource-aware hierarchical federated MTL (RHFedMTL) solution to meet the task heterogeneity corresponding to different non-IID (independent and identically distributed) training datasets. Specifically, a primal-dual method has been leveraged to effectively transform the coupled MTL into some local optimization sub-problems within BSs. Therefore, it enables solving different tasks within a BS and aggregating the multi-task result in the cloud without uploading the raw data. Furthermore, compared with existing methods that reduce resource costs by simply changing the aggregation frequency, we dive into the intricate relationship between resource consumption and learning accuracy, and develop a resource-aware learning strategy for adjusting the iteration number on local terminals and BSs to meet the resource budget. Extensive simulation results demonstrate the effectiveness and superiority of RHFedMTL in terms of improving the learning accuracy and boosting the convergence rate.

The surging demand for high-definition video streaming services and large neural network models implies a tremendous explosion of Internet traffic. To mitigate the traffic pressure, architectures with in-network storage have been proposed to cache popular contents at devices in closer proximity to users. Correspondingly, in order to maximize caching utilization, it becomes essential to devise an effective popularity prediction method. In that regard, predicting popularity with dynamic graph neural network (DGNN) models achieves remarkable performance. However, DGNN models still suffer from tackling sparse datasets where most users are inactive. Therefore, we propose a reformative temporal graph network, named semantics-enhanced temporal graph network (STGN), which attaches extra semantic information into the user-content bipartite graph and could better leverage implicit relationships behind the superficial topology structure. On top of that, we customize its temporal and structural learning modules to further boost the prediction performance. Specifically, in order to efficiently aggregate the diversified semantics that a content might possess, we design a user-specific attention (UsAttn) mechanism for the temporal learning. Unlike the attention mechanism that only analyzes the influence of genres on content, UsAttn also considers the attraction of semantic information to a specific user. Meanwhile, as for the structural learning, we introduce the concept of positional encoding into our attention-based graph learning and novelly adopt a semantic positional encoding (SPE) function, which effectively boost the performance of lightweight algorithms. Finally, extensive simulations verify the superiority of our models and demonstrate their effectiveness in content caching.

The enormous amount of network equipment and users implies a tremendous growth of Internet traffic for multi-media services. To mitigate the traffic pressure, architectures with in-network storage have been proposed to cache popular content at devices in close proximity to users in order to decrease the number of backhaul hops. Meanwhile, the reduced transmission distance also contributes to energy saving. However, due to limited storage, only a fraction of the content can be cached, while caching the most popular content is cost-effective. Correspondingly, it becomes essential to devise an effective popularity prediction method. In this regard, some existing efforts manifest the effectiveness of dynamic graph neural network (DGNN) models, but it remains challenging to tackle sparse datasets. Herein, we first propose a reformative temporal graph network, named STGN, to address the challenge and improve prediction performance. Specifically, the STGN model leverages extra semantic messages to help establish implicit paths within the sparse interaction graph and enhance the temporal and structural learning of a DGNN model. Furthermore, we devise a user-specific attention mechanism to aggregate various semantics in a fine-grained manner. Finally, extensive simulations verify the superiority of our STGN models and demonstrate the potential in terms of energy-saving.

With the continuous functional enhancement of network services, a service usually presents a directed acyclic graphic (DAG) structure. This paper models the DAG task scheduling problem as a multi-objective optimization problem to balance the task execution efficiency, network traffic, and system load balance in componentized task deployment. To produce an instant decision, we propose the Cloud-edge Collaborative Task Scheduling (CCTS) Algorithm based on hybrid reward architecture deep reinforcement learning (DRL). Specifically, to reduce the redundancy of the state space of the Markov decision process, we use directed graph convolution networks and graph convolution networks (GCN) to embed the directed task graph and undirected network graph, respectively. Simulation results show that the proposed method outperforms the compared convolutional neural networks and GCN-based DRL schemes in reducing the system latency, energy cost, network traffic, and load balance.

Multi-Agent Reinforcement Learning (MARL) has reached astonishing achievements in various fields such as the traffic control of vehicles in a wireless connected environment. In MARL, how to effectively decompose a global feedback into the relative contributions of individual agents belongs to one of the most fundamental problems. However, the volatility of the environment (e.g., the vehicle movement and wireless disturbance) could significantly shape the time-varying topological relationships among agents, thus making the Value Decomposition (VD) challenging. Therefore, in order to cope with this annoying volatility, it becomes imperative to design a dynamic VD framework. Hence, in this paper, we propose a novel Stochastic VMIX (SVMIX) methodology by embedding the dynamic topological features into the VD and incorporating the corresponding components into a multi-agent actor-critic architecture. In particular, the Stochastic Graph Neural Network (SGNN) is leveraged to effectively extract underlying dynamics embedded in topological features and improve the flexibility of VD against the environment volatility. Finally, the superiority of SVMIX is verified through extensive simulations.

Fine-grained service offloading in collaborative edge computing can make full use of the limited resource of edge nodes to achieve efficient parallel computing. It is imperative to select appropriate edge nodes for the subtask offloading in order to ensure the network’s load balance. However, there is a lack of research on computing offloading of end-to-end fine-grained services, and existing node selection algorithms can only be used in small-scale scenarios or networks with a fixed number of nodes. In this paper, we construct an end-to-end fine-grained computing offloading model, with load balancing as the optimization goal. Especially, a deep graph matching method, based on graph neural networks, is used for offloading node selection. It can be applied to dynamic and large-scale scenarios with strong generalization capability and fast execution speed. Compared with baseline algorithms, it greatly reduces the network load imbalance degree while ensuring a high acceptance ratio of services and meeting delay, location and resource constraints.

Knowledge graph completion (KGC) aims to perform link prediction to fill lost relations between entities by knowledge graph embedding (KGE). Translating embedding, as an efficient embedding method in KGE, is widely applied in numerous recent KGC models. However, these translating models may lack the ability to express various relation patterns and mapping properties for knowledge graphs (KGs). In this paper, a simple and well-performed translating model named TransC is proposed to express different relations. A multi-channel mechanism is defined firstly to constrain translating embeddings. Then a relation-aware transfer function is designed to break the expressive restriction and map triplets involving the same relation into a corresponding plane. We also mathematically prove that TransC is capable of expressing four popular relation patterns and all mapping properties. Finally, experimental results illustrate that TransC can efficiently represent the different relation patterns and properties and achieve better performance than state-of-the-art translating models.

Federated learning is an effective way to enable artificial intelligence over massive distributed nodes with security and communication efficiency. Some previous works primarily focus on learning a single global model for a unique task across the network, which is less competent to handle multi-task scenarios with stragglers and fault, after adopting the general gradient update methods in a federated environment. Others aim to learn a distinct model for each node, which is expensive in terms of the computation and communication cost. Using hierarchical network to reduce communication cost is becoming a new candidate. Thus, we propose a primal-and-dual method-based hierarchical federated multi-task learning system, supported with HFedMTL algorithm that allows massive nodes from distributed areas to join in the federated multi-task learning process. Empirical experiments verify the analysis and demonstrate the benefits of improving the learning performance and convergence rate.

In this paper, we aim to propose a novel transmission control protocol (TCP) congestion control method from a cross-layer-based perspective and present a deep reinforcement learning (DRL)-driven method called DRL-3R (DRL for congestion control with Radio access network information and Reward Redistribution) so as to learn the TCP congestion control policy in a superior manner. In particular, we incorporate the RAN information to timely grasp the dynamics of RAN, and empower DRL to learn from the delayed RAN information feedback potentially induced by several consecutive actions. Meanwhile, we relax the implicit assumption (that the feedback to one specific action returns at a round-trip-time (RTT) after the action is applied) in previous researches, by redistributing the rewards and evaluating the merits of actions more accurately. Experiment results show that besides maintaining a reasonable fairness, DRL-3R significantly outperforms classical congestion control methods (e.g., TCP Reno, Westwood, Cubic, BBR and DRL-CC) on network utility by achieving a higher throughput while reducing delay in various network environments.

Abstract Unmanned aerial vehicles (UAVs) are envisioned to provide a variety of sensing applications in the cellular systems, which is known as the cellular Internet of UAVs. This paper considers a cellular Internet of UAVs, in which the sensory data collected by UAVs can be transmitted to the base station by cellular communications, or to the corresponding mobile devices directly by overlaying UAV‐to‐device communications. The authors first propose a joint sensing and transmission protocol to schedule UAV sensing and transmission, and then investigate the energy utility (EU) maximisation problem by jointly optimising the trajectory, speed, and transmit power of UAVs in the system. As this problem is non‐convex and difficult to be solved, the authors decouple it into three subproblems , i.e. trajectory optimisation subproblem, speed optimisation subproblem, and transmit power optimisation subproblem. Then, the authors propose a joint trajectory, speed, and transmit power optimisation algorithm to obtain the suboptimal solution of the original problem. Simulation results show that the authors' proposed algorithm can efficiently improve the EU in the system as compared with other benchmark schemes.

In recent years, unmanned aerial vehicles (UAVs) have unlocked numerous sensing applications, which are expected to add billions of dollars to the world economy in the next decade. To further improve the Quality-of-Service in these applications, the 3rd Generation Partnership Project has considered the use of terrestrial cellular networks to support UAV sensing services, also known as the cellular Internet of UAVs. In this paper, we consider a cellular Internet of UAVs, where the sensory data can be transmitted either to the base station via cellular links, or to the mobile devices by underlay UAV-to-Device (U2D) communications. To evaluate the freshness of the sensory data, the concept of age of information (AoI) is adopted, in which a lower AoI implies fresher data. Since UAVs' AoIs are determined by their trajectories during sensing and transmission, we investigate the AoI minimization problem for UAVs by designing their trajectories. This problem is a Markov decision problem with an infinite state-action space, and thus we utilize multi-agent deep reinforcement learning to approximate the state-action space. Then, we propose a multi-UAV trajectory design algorithm to solve this problem. Simulation results show that our proposed algorithm can achieve a lower AoI than a greedy algorithm, policy gradient algorithm, and overlay U2D scheme.

In this paper, we consider a cellular Internet of UAVs, where the sensory data can be transmitted either to the base station via cellular links, or to the mobile devices by underlay UAV-to-Device communications. To evaluate the freshness of the sensory data, the age of information (AoI) is adopted, in which a lower AoI implies fresher data. Since UAVs' AoIs are determined by their trajectories during sensing and transmission, we aim to minimize the AoIs of UAVs by designing their trajectories. This problem is a Markov decision problem with an infinite state-action space, and thus, we propose a multi-UAV trajectory design algorithm by leveraging multi-agent deep reinforcement learning to solve it. Simulation results show that our proposed algorithm outperforms both a greedy algorithm and a policy gradient algorithm.

In the current unmanned aircraft systems (UASs) for sensing services, unmanned aerial vehicles (UAVs) transmit their sensory data to terrestrial mobile devices over the unlicensed spectrum. However, the interference from surrounding terminals is uncontrollable due to the opportunistic channel access. In this paper, we consider a cellular Internet of UAVs to guarantee the Quality-of-Service (QoS), where the sensory data can be transmitted to the mobile devices either by UAV-to-Device (U2D) communications over cellular networks, or directly through the base station (BS). Since UAVs' sensing and transmission may influence their trajectories, we study the trajectory design problem for UAVs in consideration of their sensing and transmission. This is a Markov decision problem (MDP) with a large state-action space, and thus, we utilize multi-agent deep reinforcement learning (DRL) to approximate the state-action space, and then propose a multi-UAV trajectory design algorithm to solve this problem. Simulation results show that our proposed algorithm can achieve a higher total utility than policy gradient algorithm and single-agent algorithm.

In this paper, we consider a cellular Internet of unmanned aerial vehicles (UAVs) where the sensory data can be transmitted to the mobile devices directly by overlaying UAV-to-Device (U2D) communications, or through the base station (BS) by cellular communications. Since the transmission modes of UAVs may influence their trajectories, we study the trajectory design problem for UAVs aiming to maximize the total utility in consideration of their transmission modes. This problem is a Markov decision problem (MDP) with a large state-action space, and thus, we propose a multi-UAV trajectory design algorithm using multi- agent deep reinforcement learning (DRL) to solve this problem. Simulation results show that our proposed algorithm can achieve a higher total utility than the single-agent method.

In this paper, we consider a device-to-device (D2D) communications underlaying cellular network where Long Term Evolution (LTE) and D2D users are allowed to communicate over both licensed and unlicensed bands for spectrum efficiency improvement. LTE users utilize spectrum orthogonally and share it with D2D users. To maximize the total throughput of this D2D system, we leverage stochastic geometry to derive the throughput for each kind of users by modeling the deployment of users as Poisson point processes (PPPs), and investigate the mode selection problem for D2D users. Since the problem is NP-hard, we propose a sequential quadratic programming (SQP) based algorithm to obtain the corresponding suboptimal solutions. Theoretically, we evaluate the system performance by analyzing the throughput regions. Simulation results validate the accuracy of the geometric analysis and verify the effectiveness of the proposed algorithm.

The DC power packet transmission contributes to a reliable integration of distributed energy resources (DERs) in power grids and reduces the load fluctuation. In this paper, we consider a DC packetized-power microgrid for the integration of DERs and propose a power packet dispatching protocol to regulate the peer to peer power interchange within the microgrid. We formulate the joint subscriber matching and energy allocation problem to optimize the subscribers' benefits, which is proved to be NP-hard. Based on the matching theory, we associate the problem equivalent to a many-to-many matching problem and design a two-sided matching algorithm to solve it. We then design a graph coloring based algorithm to schedule the energy transmissions of the matched energy subscribers. Simulation results validate the effectiveness of the proposed protocol in achieving a steady and efficient microgrid power dispatching.

In this paper, we consider device-to-device (D2D) communications as an underlay to cellular networks over both licensed and unlicensed spectrums, where long-term evolution (LTE) users utilize the spectrum orthogonally while D2D users share the spectrum with LTE users. In the system, each LTE and D2D user can access the licensed or unlicensed band for communications. To maximize the total throughput of the system, we leverage stochastic geometry to derive the throughput for each kind of user by modeling the deployment of users as Poisson point processes (PPPs), and investigate the spectrum access problem for these users. Since the problem is NP-hard, we propose a sequential quadratic programming (SQP)-based algorithm to obtain the corresponding suboptimal solutions. Theoretically, we evaluate the system performance by analyzing the throughput regions. Simulation results validate the accuracy of the geometric analysis and verify the effectiveness of the proposed algorithm.

In this paper, we study an X-duplex relay system with one source, one amplify-and-forward relay, and one destination, where the relay is equipped with a shared antenna and two radio frequency (RF) chains used for transmission or reception. X-duplex relay can adaptively configure the connection between its RF chains and antenna to operate in either half-duplex (HD) or full-duplex (FD) mode, according to the instantaneous channel conditions. We first derive the distribution of the signal to interference plus noise ratio, based on which we then analyze the outage probability, average symbol error rate (SER), and average sum rate. We also investigate the X-duplex relay with power allocation and derive the lower bound and upper bound of the corresponding outage probability. Both analytical and simulated results show that the X-duplex relay achieves a better performance over pure FD and HD schemes in terms of SER, outage probability and average sum rate, and the performance floor caused by the residual self interference can be eliminated using flexible RF chain configurations.

In this paper, a novel X-Duplex relay scheme with one source, one amplify-and-forward (AF) relay and one destination is proposed. The relay is equipped with a shared antenna and two radio frequency (RF) chains used for transmission or reception. The proposed scheme can be reduced to either full-duplex (FD) or half-duplex (HD) with different RF chain configurations. In the proposed scheme, relay adaptively configures the connection between its RF chains and the antenna to optimise the end-to-end system performance according to the instantaneous channel conditions. In this paper, we analyze the system overall performances based on the distribution of the signal to interference plus noise ratio (SINR) of the hybrid mode, including outage probability and average sum rate. Monte-Carlo simulations are used to validate the analytical expressions. Results show that the X-Duplex relay achieves a lower outage probability and a higher average sum rate compared to FD and HD schemes.

Driven by both personal and commercial interests, fast popular content delivery, as one of the key services offered by vehicular ad-hoc networks (VANETs), has recently received considerable attention. Most existing work mainly focuses on the resource allocation such as transmit power or subcarrier assignment from the on-board units (OBUs) to the roadside units (RSUs). However, due to the limited backhaul capacity, great efforts still need to be taken for delivering large-size files such as videos and music to the high speed vehicles. Motivated by the recent work of pre-storing files in the cell-edge base stations, in this paper, we address the efficient content delivery problems in VANET by caching popular files in the RSUs with large storage capacity. The main objective is to minimize the average time that an OBU downloads a file. We propose three algorithms of allocating files to RSUs, in the optimal, sub-optimal, and greedy ways respectively, where the first one can achieve the best performance, and the greedy one has the lowest complexity. We also analyze the average downloading time performance in terms of the number of RSUs, storage capacity, and vehicle speed. Simulation results indicate that the proposed RSU caching methods can significantly reduce the file-downloading time, and thus, improve the content delivery efficiency.

Machine‐to‐machine (M2M) communication plays an important role in various kinds of intelligent networks. In this study, a hybrid cooperation scheme for data collection in hierarchical smart building networks (SBN) is proposed under the framework of M2M communications. The hierarchical network structure means that the data collection process is carried out via multi‐layer communications. In the first layer, smart metres organise themselves into clusters and send information to the cluster‐heads. Then all cluster‐heads forward the received information to the base station automatically in the second layer. In particular, the roles of cluster‐head can be acted by either fixed nodes or user terminals in the building, and this endow a hybrid cooperation mode to the data collection process. To construct the network structure and utilise the resources efficiently, the authors first provide some theoretical analysis on the influence of network structure and bandwidth constraints. Then a distributed scheme for joint structure formation and subband allocation is proposed based on coalitional game theory. Furthermore, for the feasibility of this scheme in practical applications, some improvements of the proposed scheme have also been made at last. The advantages of the proposed scheme are verified by simulation results.

In this study, the authors introduce adaptive modulation and coding in a two‐way amplify‐and‐forward (AF) relay network to improve the system performance. They consider a cooperative system where two user nodes exchange information with the assistance of multiple two‐way AF relays. In the proposed scheme, the user nodes adaptively choose the appropriate modulation and coding scheme to ensure that the frame error rate (FER) satisfies the system requirement; all relays are utilised to forward the received signals to the user terminals. Furthermore, they provide a better approximation of the cumulative distribution function of the destination signal‐to‐noise ratio; and thus, derive more accurate expressions including average spectral efficiency and average FER in closed‐form, over Rayleigh fading channels. The theoretical analysis is verified by numerical results.

It is well recognized that channel state information (CSI) feedback plays a key role in the performance of closed-loop wireless networks. However, most work on energy efficiency (EE) typically studies this problem from the downlink data transmission point of view. In this paper, we propose an alternative approach to investigate the EE for wireless communication networks through controlling the channel state information (CSI) in the feedback link, in which a number of multiple-antenna mobile transmitters exchange information with their corresponding mobile receivers using linear precoding for interference reduction. Specially, we formulate this EE maximization problem in the analytical setting of a game theoretic framework, and propose a two-level Stackelberg-type CSI feedback control game (SCFC) to balance the bandwidth and power consumptions in a distributed manner. The existence of the equilibriums of such games is proved, and the convergence behavior is investigated. Simulation results show that by adjusting the pricing factor, the proposed distributed SCFC game effectively improves the EE performance.

A new noise variance based reduced maximum likelihood decision feedback equalization (ML-DFE) algorithm has been developed. This algorithm reduces the calculation complexity by exploring the intrinsic statistical properties layer by layer. Through setting layered thresholds, part of the nodes in the searching process will be cut by comparing with the thresholds. Simulation results show that the complexity drops lots while the performance drops small.

In this study, the superimposed training strategy is introduced into orthogonal frequency division multiplexing-modulated amplify-and-forward two-way relay network (TWRN) to perform two-hop transmission-compatible individual channel estimation. Through the superposition of an additional training vector at the relay under power allocation, the separated source–relay channel information can be directly obtained at the destination and then used to estimate the channels. The closed-form Bayesian Cramér-Rao lower bound (CRLB) is derived for the estimation of block-fading frequency-selective channels with random channel parameters, and orthogonal training vectors from the two source nodes are required to keep the Bayesian CRLB simple because of the self-interference in the TWRN. A set of optimal training vectors designed from the Bayesian CRLB are applied in an iterative linear minimum mean-square-error channel estimation algorithm, and the mean-square-error performance is provided to verify the Bayesian CRLB results.

The use of high level trellis-coded modulation such as TCM 16-QAM, with equalization, diversity, interleaving and slow frequency hopping is investigated for wideband mobile communications. The symbol error rate (SER) performance of those systems in presence of intersymbol interference (ISI) and Rayleigh fading is evaluated by theoretical analysis and computer simulation. Several schemes of combining equalizers with diversity are evaluated by computer simulations. It shows that by choosing a proper interleaving size, number of the frequency hopping and the combining scheme of equalization and space diversity, the proposed system gives considerable performance improvement and can cope with the frequency selective multipath fading in wideband mobile communication systems.

The paper presents a new reduced-state soft decision feedback Viterbi equalizer (RSSDFVE) with a channel estimator and predictor. A multi-ray Rayleigh fading channel model with a Doppler frequency shift is used in the simulation. For fast convergence, a channel estimator with fast start-up is proposed. The channel estimator obtains the sampled channel impulse response (CIR) from the training sequence and updates the RSSDFVE during the bursts in order to track the changes of the channel. According to the simulation results, the proposed RSSDFVE gives much better performance than the DFE for severe Rayleigh multipath fading channels with much less implementation complexity than MLSE. The fast start-up (FS) channel estimator gives a faster convergence than a Kalman channel estimator and can achieve the same performance as the Kalman estimator with a shorter training sequence.>