With growing complexity in urban transportation systems, multimodal traffic flow prediction is essential for coordinating multiple transport modes, optimizing resource allocation, and enhancing urban mobility efficiency. To improve prediction accuracy, it is essential to model cross-modal correlations effectively. However, existing methods in cross-modal correlation modeling overlook two critical issues: 1) Existing cross-modal graph construction methods fail to uncover the latent relationships between nodes based on the travel patterns across different modes; 2) Existing cross-modal feature extraction methods are limited to a single scale and lack comprehensive spatiotemporal representations, resulting in incomplete modeling of cross-modal dependencies. To enhance cross-modal correlation capture, we propose a novel framework named Multi-Scale Cross-Modal Spatiotemporal Enhanced Network (MCSTEN). Specifically, we design a cross-modal graph generator that utilizes tensor decomposition to uncover latent node correlations from peak-time asynchrony features across modes and constructs a cross-modal graph with rich information. Subsequently, we introduce a multi-scale feature extraction block equipped with a scale transformation mechanism to enhance the modeling of cross-modal correlations at different temporal and spatial scales. Finally, an attention-based fusion strategy integrates these multi-scale correlations, adaptively balancing the contribution of each scale. Experimental results on seven real-world multimodal transportation datasets demonstrate that MCSTEN significantly outperforms state-of-the-art methods in prediction accuracy.

Long-range (LoRa) technology has emerged as a promising solution for Internet of Things applications due to its low power consumption and long communication range. However, its pure ALOHA-based MAC layer leads to severe packet collisions as the network scale expands, significantly degrading the system energy efficiency (EE). While careful allocation of transmission parameters such as channel (CH), transmission power (TP), and spreading factor (SF) could mitigate this issue, the complex interference patterns in multi-gateway scenarios and the time-consuming nature of EE evaluation pose significant challenges. Therefore, we propose an analytical model to calculate the system EE while fully considering the impacts of multiple gateways, duty cycling, quasi-orthogonal SFs and capture effects. Based on this model, we formulate a joint CH, SF, and TP allocation problem to optimize the system EE. To solve this NP-hard optimization problem, we decompose it into CH assignment and SF/TP assignment subproblems. A two-phase optimization framework is then designed. In the first phase, a matching-based algorithm is designed for CH assignment. In the second phase, a multi-agent reinforcement learning approach that incorporates a two-stage attention mechanism and graph convolutional networks is proposed for SF/TP assignment, which effectively captures and weights inter-ED interactions in multi-gateway scenarios. Simulation results indicate that the proposed approach well-suited for complex multi-GW LoRa network topologies and outperforms state-of-the-art algorithms.

Multivariate long-term time series forecasting is pivotal across numerous domains, yet precise predictions require a differentiated assessment of historical time segments due to their varying influence on future trends. Patch-based Transformer frameworks show promise for capturing local temporal patterns. However, they face limitations with static patching, which disrupts temporal continuity, fails to adapt to shifts between periodic and volatile patterns, and overlooks dynamic interactions between time segments and variables. To address these limitations, we propose Entropy-Aware Patch Transformer (EAPformer) which dynamically segments time series for differentiated assessments of historical patterns. Specifically, we overcome static patching limitations by leveraging temporal entropy to dynamically adjust patch boundaries through a two-stage policy, achieving interpretable and context-sensitive segmentation. Subsequently, we adapt EAPformer to periodic and volatile dynamics by employing entropy-aware segmentation that captures distinct temporal patterns across diverse segments. Finally, we further capture dynamic interactions across time segments and variables by introducing a multi-dimensional dependency learning architecture. Additionally, a gated fusion mechanism integrates local and global patterns, enhancing robustness. Extensive experiments on eight public benchmarks demonstrate that EAPformer outperforms state-of-the-art models, achieving superior accuracy across all metrics.

In this paper, we consider a large-scale Wi-Fi HaLow heterogeneous network, where numerous stations (STAs) are distributed around an access point (AP) to collect and transmit data using the restricted access window (RAW) mechanism. The AP manages channel access by adjusting and broadcasting RAW Parameter Set (RPS) and grouping messages, which include the duration and slot allocation for each RAW group. We aim to maximize the overall network throughput while ensuring fairness among STAs. Traditional methods struggle with real-time RAW optimization in practical networks. To overcome this challenge, we first divide the RAW groups heuristically according to the STAs' task type and formulate the throughput optimization problem regarding various RPS. Then, we construct a virtual twin network environment to estimate the throughput performance used for RAW optimization. Specifically, the twin environment is built on neural networks and trained by both synthetic data generated from the classic Markov model and the NS-3 simulator. Given the throughput estimate, we devise the reward function of the proximal policy optimization (PPO) algorithm to adapt the optimal RPS, without frequent interaction with the real network environment. Numerical results indicate that the twin-enhanced PPO (TE-PPO) algorithm achieves a comparable performance with the classic PPO algorithm built on the real trace of the NS3 simulator. Particularly, TE-PPO can reduce the time overhead for RPS adaptation to 1/180.

This paper focuses on secure communications in UAV-assisted wireless sensing systems in the presence of a mobile UAV eavesdropper (MUE), without prior information about the traffic demands of ground users (GUs) in the sensing environment. To maximize secrecy performance, we propose a digital-twin enhanced proximal policy optimization (DTEPPO) algorithm that optimizes the GUs' sensing scheduling and the UAVs' trajectory planning and network formation. Unlike traditional reinforcement learning methods that require complete environment-interacted information, we build a digital twin (DT) model using Gaussian process regression (GPR) based on the historical sensing information collected by the UAVs. Moreover, DT model can be dynamically updated by exploiting the PPO algorithm to lean the UAVs' interactions with the environment, continuously enhancing its accuracy and providing a reliable virtual learning environment for fast network adaptation. Numerical simulations demonstrate that the proposed DTEPPO algorithm achieves rapid convergence and improved secure throughput with reduced communication overhead compared to conventional approaches. Moreover, the proposed framework enables simultaneous learning and modeling (SLAM) in an unknown environment, providing a general framework to solve complex network control problems in wireless and mobile systems with high costs of environmental interactions.

The imbalance in network resource allocation and service quality across cities remains a critical issue that demands effective solutions. Accurate city-scale cellular traffic prediction plays a vital role in addressing this challenge. However, existing methods heavily rely on abundant data, limiting their applicability in data-scarce cities. To overcome this limitation, we propose a two-stage Cross-city Transfer Learning framework for cellular traffic Prediction (CTLP). In the first stage, a spatiotemporal feature concatenation network is proposed. This network captures the dynamics of cellular traffic derived from a data-rich city and a data-scarce city with attention mechanisms and then aggregates these dynamics with a CNN. In the second stage, parameters of the first-stage network are transferred from the data-rich cities to the data-scarce cities for addressing the challenge of cellular traffic heterogeneity in cross-city transfer and enabling more accurate predictions. Experiments on a realworld cross-city cellular traffic dataset demonstrate that CTLP significantly outperforms existing methods, effectively solving the problem of cellular traffic prediction in data-scarce cities.

Clustering user devices (UDs) with similar traffic flows enables more effective transmission scheduling and resource allocation in wireless networks, especially for Internet of Things (IoT) with dominant demands for machine-to-machine (M2M) data communications. In this paper, we propose an experience-driven spatial-temporal graph attention network (Exp-STGAN) for UDs’ clustering, by exploiting the spatial-temporal correlations from UDs’ historical traffic flows. Considering the UDs’ heterogeneity, we first characterize each UD’s out-going traffic flows by a dynamic radiation pattern, which reflects the UD’s spatial distribution of traffic demands and its targeted receivers in different directions. We aim to explore UDs’ clustering based on traffic flows’ radiation patterns and propose a spatial-temporal graph attention to aggregate information from different UDs correlated in space and time domains. Without true labels for the UDs’ clustering, we formulate a flow similarity metric based on Kullback-Leibler (KL) divergence to quantify the clustering performance. Moreover, to improve the learning efficiency, we integrate salient human experience into the graph attention module and also continuously update the experience during the training process. Experiments demonstrate that the Exp-STGAN framework can effectively cluster similar UDs by their dynamic traffic flows, highlighting the potential for flow-aware network performance maximization in large-scale IoT systems.

In mobile edge computing systems, base stations (BSs) equipped with edge servers can provide computing services to users to reduce their task execution time. However, there is always a conflict of interest between the BS and users. The BS prices the service programs on the basis of the user demand to maximize its own profit, whereas the users determine their offloading strategies based on the prices to minimize their costs. Moreover, service programs need to be precached to meet immediate computing needs. Due to the limited caching capacity and variations in service program popularity, the BS must dynamically select which service programs to cache. Since service caching and pricing have different needs for adjustment time granularities, we propose a two-time scale framework to jointly optimize service caching, pricing and task offloading. For the large time scale, we propose a game-nested deep reinforcement learning algorithm to dynamically adjust service caching according to the estimated popularity information. For the small time scale, by modeling the interaction between the BS and users as a two-stage game, we prove the existence of the equilibrium under incomplete information and then derive the optimal pricing and offloading strategies. Extensive simulations based on a real-world dataset demonstrate the efficiency of the proposed approach.

In this paper, we investigate hybrid passive and active intelligent reflecting surface (IRS)-assisted secure non-orthogonal multiple access (NOMA) networks. Multiple users concurrently transmit sensitive data to an access point (AP) in the presence of an eavesdropper (Eve). The hybrid IRS is employed to enhance the NOMA users’ sum rates while simultaneously performing jamming beamforming against the Eve by optimizing the communication channels of NOMA users and injecting controllable noise into the Eve’s channel. We formulate a sum secrecy rate maximization problem by jointly optimizing the users’ scheduling policy, the hybrid IRS’s working mode and beamforming, and the AP’s receiving beamforming. To address combinatorial user scheduling and high-dimensional beamforming design, we develop a dual-cycling deep reinforcement learning (DRL) framework. We first determine the NOMA users’ scheduling strategy and the hybrid IRS’s working mode using a proximal policy optimization (PPO)-based learning algorithm. Then, we optimize the AP’s receiving beamforming and hybrid IRS’s beamforming strategies using an alternating optimization (AO) algorithm. The joint beamforming optimization can significantly enhance the DRL’s learning efficiency by limiting its action space. Moreover, we propose a lightweight two-phase algorithm with approximation techniques to reduce computational complexity by eliminating double-nested loops in AO, while maintaining secrecy performance close to optimum. Numerical results demonstrate that the proposed dual-cycling DRL scheme achieves 54.85% gains in the secrecy rate compared to traditional DRL schemes.

Mobile crowdsensing (MCS) has emerged as a prominent trend across various domains. However, ensuring the quality of the sensing data submitted by mobile users (MUs) remains a complex and challenging problem. To address this challenge, an advanced method is needed to detect low-quality sensing data and identify malicious MUs that may disrupt the normal operations of an MCS system. Therefore, this article proposes a prediction- and reputation-based truth discovery (PRBTD) framework, which can separate low-quality data from high-quality data in sensing tasks. First, we apply a correlation-focused spatio-temporal Transformer network that learns from the historical sensing data and predicts the ground truth of the data submitted by MUs. However, due to the noise in historical data for training and the bursty values within sensing data, the prediction results can be inaccurate. To address this issue, we use the implications among the sensing data, which are learned from the prediction results but are stable and less affected by inaccurate predictions, to evaluate the quality of the data. Finally, we design a reputation-based truth discovery (TD) module for identifying low-quality data with their implications. Given the sensing data submitted by MUs, PRBTD can eliminate the data with heavy noise and identify malicious MUs with high accuracy. Extensive experimental results demonstrate that the PRBTD method outperforms existing methods in terms of identification accuracy and data quality enhancement.

In this paper, we employ an unmanned aerial vehicle (UAV) to ensurethe freshness of sensing data, as measured by the age of information (AoI), in Internet of Things (IoT) networks. Specifically, the UAV switches between flying and hovering modes to collect data from widely distributed IoT devices. UAV trajectory planning, which determines the times and moments of data collection, is vital for optimizing the system AoI. Considering the limited UAV onboard energy and mission duration, AoI-optimal trajectory planning is formulated as a mixed-integer nonlinear programming (MINLP) problem. We first decompose the MINLP problem into two subproblems: a UAV time scheduling subproblem and a UAV path planning subproblem. Then, we propose a learning-based iterative (LBI) algorithm that consists of two modules: a successive convex approximation (SCA)-based module for solving the time scheduling subproblem, and a hierarchical asynchronous advantage actor-critic (A3C) module for addressing the path planning subproblem. The numerical results verify that the proposed LBI algorithm outperforms typical baselines in terms of the AoI performance.

In this paper, we consider an aerial reconfigurable intelligent surface (ARIS)-assisted wireless network, where multiple unmanned aerial vehicles (UAVs) collect data from ground users (GUs) by using the non-orthogonal multiple access (NOMA) method. The ARIS provides enhanced channel controllability to improve the NOMA transmissions and reduce the co-channel interference among UAVs. We also propose a novel dual-mode switching scheme, where each UAV equipped with both an ARIS and a radio frequency (RF) transceiver can adaptively perform passive reflection or active transmission. We aim to maximize the overall network throughput by jointly optimizing the UAVs’ trajectory planning and operating modes, the ARIS’s passive beamforming, and the GUs’ transmission control strategies. We propose an optimization-driven hierarchical deep reinforcement learning (O-HDRL) method to decompose it into a series of subproblems. Specifically, the multi-agent deep deterministic policy gradient (MADDPG) adjusts the UAVs’ trajectory planning and mode switching strategies, while the passive beamforming and transmission control strategies are tackled by the optimization methods. Numerical results reveal that the O-HDRL efficiently improves the learning stability and reward performance compared to the benchmark methods. Meanwhile, the dual-mode switching scheme is verified to achieve a higher throughput performance compared to the fixed ARIS scheme.

In mobile edge computing systems, base station (BS) equipped with edge servers can provide computing services to users to reduce their task durations. The BS prices the service programs based on user demand to maximize its own profits. Additionally, due to limited caching capacity and variations in service programs popularity, the BS has to dynamically select which service programs to cache. To address the conflict between high profits requirement and system instability, we propose a two time-scale framework to optimize service caching, pricing and task offloading. Under the small time scale, by modeling the interaction between the BS and users as a two-stage game, we derive the optimal offloading strategy and pricing algorithm. Then, we deduce the existence of equilibrium points. Under the large time scale, we propose a game-nested deep reinforcement learning algorithm to dynamically adjust the service caching according to estimated popularity information. Extensive data based simulations demonstrate the efficiency of the proposed approach.

LoRa is considered one of the most promising low-power wide-area techniques. Given that end devices (EDs) are typically battery-powered, energy efficiency (EE) is a critical factor to consider. In this paper, we aim to improve the system EE of the LoRa network by jointly allocating transmission parameters such as the channel (CH), transmission power (TP) and spreading factor (SF) for each ED. Owing to the low duty cycle and sporadic traffic of LoRa networks, evaluating the system EE under various parameter settings proves to be time-consuming. Consequently, we propose an analytical model aimed at calculating the system EE while fully considering the impact of multiple gateways, quasi-orthogonal SFs and capture effects. On this basis, we investigate a joint CH, SF and TP allocation problem to optimize the system EE for uplink transmissions. Given the NP-hard complexity of the problem, the original problem is decomposed into two subproblems: CH assignment and SF/TP assignment. First, a matching-based algorithm is introduced to tackle the CH assignment subproblem. Then, an attention-based multiagent reinforcement learning technique is employed to address the SF/TP assignment subproblem for EDs allocated to the same CH. The simulation outcomes indicate that the proposed approach converges quickly and obtains significantly better system EE than baseline algorithms.

Physical layer security issues have attracted significant attention in wireless networks to protect information leakage from illegitimate eavesdroppers. In this paper, we focus on an intelligent reflecting surface (IRS)-assisted secure non-orthogonal multiple access (NOMA) uplink system. Multiple users intend to transmit sensitive data to an access point (AP) considering the existence of a nearby eavesdropper (Eve). The IRS can be used to enhance the NOMA users’ sum rates while concurrently weakening the Eve’s channel condition, suppressing information leakage to the Eve without resorting to a cooperative jammer or the power injection of artificial noise in the system. The users’ scheduling, the IRS’s passive beamforming, and the AP’s receive beamforming are jointly optimized to maximize the secure rate of the IRS-assisted NOMA system. We develop a hierarchical deep reinforcement learning (DRL) framework to iteratively search for an optimal solution considering the combinatorial nature of the NOMA users’ scheduling and the high-dimensional beamforming design. Firstly, we search for the NOMA users’ scheduling strategy by using the PPO-based DRL algorithm. Given the NOMA scheduling strategy, we then optimize the active and passive beamforming strategies by the alternating optimization (AO) algorithm. The inner optimization helps evaluate the quality of the scheduling strategy and thus guides the outer-PPO algorithm to update a better scheduling strategy. Simulation results demonstrate the superiority of the proposed scheme over existing benchmarks, resulting in significant gains in secure rate.

Mobile crowdsensing (MCS) has emerged as a prominent trend across various domains. However, ensuring the quality of the sensing data submitted by mobile users (MUs) remains a complex and challenging problem. To address this challenge, an advanced method is required to detect low-quality sensing data and identify malicious MUs that may disrupt the normal operations of an MCS system. Therefore, this article proposes a prediction- and reputation-based truth discovery (PRBTD) framework, which can separate low-quality data from high-quality data in sensing tasks. First, we apply a correlation-focused spatio-temporal transformer network that learns from the historical sensing data to predict the ground truth of the data submitted by MUs. However, due to the noise in historical data for training and the bursty values within sensing data, the prediction results can be inaccurate. To address this, we utilize the implications among sensing data, which are extracted from the prediction results but are stable and less affected by inaccurate predictions, to evaluate the quality of data. Finally, we design a reputation-based truth discovery (TD) module for identifying low-quality data with their implications. Given sensing data submitted by MUs, PRBTD can eliminate the data with heavy noise and identify malicious MUs with high accuracy. Extensive experimental results demonstrate that PRBTD outperforms the existing methods in terms of identification accuracy and data quality enhancement.

In the rapidly evolving field of Internet of Vehicles technology, Cellular Vehicle-to-Everything (C-V2X) communication has garnered considerable attention due to its superior performance in terms of coverage, latency, and throughput. Efficient resource allocation in C-V2X networks is crucial for ensuring the transmission of safety-critical information in Vehicle-to-Vehicle (V2V) communication. In this paper, we propose Dynamic-aware Graph-based Proximal Policy Optimization (DGPPO), a novel approach that integrates Graph Neural Networks (GNNs) with Deep Reinforcement Learning to effectively tackle the challenges of joint spectrum and power allocation in C-V2X networks. The proposed approach constructs a dynamic graph representation where communication links are modeled as nodes, allowing for flexible modeling of network topology changes. Next, we utilize an adaptive GNN to extract essential low-dimensional global features from the dynamically evolving graph network topology, which is constructed based on local observations. Finally, we propose a multi-agent proximal policy optimization (MAPPO) approach that leverages these extracted features to promote cooperative behavior and adaptability among agents. The simulation results validate the effectiveness of the DGPPO algorithm, outperforming the state-of-the-art algorithms in highly dynamic vehicular network environments.

In this paper, we aim to improve the throughput capacity of a wireless powered network by allowing user devices (UDs) to adapt channel access strategies. A base station (BS) can receive data and provide RF energy for UDs simultaneously in a full duplex mode. Each UD can choose a flexible access mode to transmit its data with the non-orthogonal multiple access technique or assist the other UDs' data transmissions via backscattering when it has less urgent data demands or insufficient energy supply. We maximize the sum throughput by optimizing the UDs' channel access mode, time allocation and beamforming strategies, according to the UDs' channel conditions and energy status. Practically, maximizing throughput is a challenging problem due to the uncertain channel information, the UDs' dynamic traffic demands and limited energy storages. Therefore, we propose a hierarchical learning approach to decompose the access mode selection and transmission control in two steps. We first employ a multi-agent deep reinforcement learning approach to update each UD's channel access strategy by interacting with the uncertain network environment. Then, we efficiently update the UDs' time allocation and the BS's beamforming strategies to further enhance the sum throughput. The simulation results verify a significant improvement in terms of the throughput and the learning efficiency compared to the benchmark methods.

This paper exploits multiple unmanned aerial vehicles (UAVs) to assist energy transfer, data uploading, and transmission in wireless networks, aiming to maximize the network's energy efficiency (EE). The inherent challenge of inaccessible or energy-intensive real-time information exchanges among UAVs results in undesirable delays in acquiring global network information. Such delayed information significantly hinders the transmission control and trajectory planning of the UAV s in multi-UAV-assisted wireless networks. To address this challenge, we propose a delay-tolerant multi-agent deep reinforcement learning (DT-MADRL) algorithm to jointly optimize the UAVs' trajectories and transmission control strategies based on randomly delayed information. In particular, we integrate a delay penalty term in the reward function that forces each UAV to have more regular information exchanges with the base station (BS). This ensures that each UAV can understand the real-time network environment, thereby reducing information delay and fostering more effective multi-agent collaboration. The simulation results reveal that our proposed algorithm reduces the UAVs' average information delay by 68% and improves overall EE by 28% compared to traditional MADRL algorithms.

In this paper, we consider a multiple unmanned aerial vehicles (UAVs)-assisted wireless sensing network, where low-power ground users (GUs) periodically sense the environmental information and upload the recent sensing information to a base station (BS). The GUs firstly backscatter their information to the UAVs and then the UAVs transmit the information to the BS by the non-orthogonal multiple access (NOMA) transmissions. Our goal is to minimize the long-term age-of-information (AoI) by jointly optimizing the UAV's sensing scheduling, transmission control, and trajectories. To solve this problem, we propose the Lyapunov-driven hierarchical proximal policy optimization framework, named Lya-HPPO, to decouple the multi-stage AoI minimization problem into several control subproblems. In each control subproblem, the UAVs' sensing scheduling and transmission control are firstly determined by the outer-loop deep reinforcement learning (DRL) approach, and then the inner-loop optimization module is to update the UAVs' trajectories. Simulation results verify that the proposed Lya-HPPO framework converges very fast to a stable value and can make online decisions in real time, while guaranteeing the long-term data buffer and AoI stability.

In this paper, we employ dual-mode unmanned aerial vehicles (UAVs) equipped with both the active radio frequency (RF) module and aerial reconfigurable intelligent surface (ARIS) to assist ground users (GUs) for both the downlink energy transfer and uplink data transmission in a wireless-powered network. To maximize the GUs' minimum throughput, we propose a collaborative mode switching scheme for the dual-mode UAVs to dynamically switch between the active RF and passive ARIS modes according to the time-varying channel conditions. Besides, we jointly optimize the GUs' transmission control, the UAVs' beamforming, and the trajectory planning strategies. This optimization problem is intractable directly due to the non-convexity in both the objective and constraints. We design an iterative algorithm to first decompose the original problem into several subproblems, and then solve each subproblem individually by approximate optimization methods. Numerical results verify that the UAVs' collaborative mode switching along with their trajectories efficiently improves the transmission performance compared to the benchmarks in which both UAVs are operating in one fixed mode.

Long Range (LoRa) wireless technology, characterized by low power consumption and a long communication range, is regarded as one of the enabling technologies for the Industrial Internet of Things (IIoT). However, as the network scale increases, the energy efficiency (EE) of LoRa networks decreases sharply due to severe packet collisions. To address this issue, it is essential to appropriately assign transmission parameters such as the spreading factor and transmission power for each end device (ED). However, due to the sporadic traffic and low duty cycle of LoRa networks, evaluating the system EE performance under different parameter settings is time-consuming. Therefore, we first formulate an analytical model to calculate the system EE. On this basis, we propose a transmission parameter allocation algorithm based on multiagent reinforcement learning (MALoRa) with the aim of maximizing the system EE of LoRa networks. Notably, MALoRa employs an attention mechanism to guide each ED to better learn how much “attention” should be given to the parameter assignments for relevant EDs when seeking to improve the system EE. Simulation results demonstrate that MALoRa significantly improves the system EE compared with baseline algorithms with an acceptable degradation in packet delivery rate (PDR).

Intelligent reflecting surface (IRS) has been recently exploited as a symbiotic radio (SR) technology to improve energy and spectral efficiencies in wireless systems. In this article, we consider a symbiotic IRS-assisted mobile-edge computing (MEC) system that allows edge users to first harvest RF power from a hybrid access point (HAP) and then offload its computational workload to the MEC server associated with the HAP. We aim to minimize the HAP’s energy consumption by jointly optimizing the users’ offloading schemes, the HAP’s active beamforming, and the IRS’s passive beamforming strategies. We propose an optimization-driven hierarchical deep deterministic policy gradient (OH-DDPG) framework to decompose the energy minimization problem into the optimization and the learning subproblems, respectively. The outer loop DDPG learning method adapts the IRS’s passive beamforming strategy, while the inner loop optimization deals with the other control variables with reduced dimensionality. Moreover, to improve the learning efficiency, we extend OH-DDPG to the multiagent scenario. In particular, the HAP first estimates the users’ offloading strategy by the inner-loop optimization and shares it with all user agents. Then, each user agent refines its offloading decision using the DDPG algorithm independently. This can avoid signaling overhead among users and improve the multiuser learning efficiency. Simulation results show that the proposed OH-DDPG and the multiuser extension can achieve significant performance gains compared to the conventional model-free learning algorithms.

Timely and accurate cellular traffic prediction is difficult to achieve due to the complex spatial-temporal characteristics of cellular traffic. The latest approaches mainly aim to model local spatial-temporal dependencies of cellular traffic based on deep learning techniques but lack the consideration of diverse global spatial-temporal correlations hidden in cellular traffic. To tackle this issue, we propose a novel multi-view spatial-temporal graph network (MVSTGN), which combines attention and convolution mechanisms into traffic pattern analysis, enabling the comprehensive excavation of spatial-temporal characteristics. Specifically, the MVSTGN realizes the above statement from three spatial-temporal views: 1) From a global spatial view, two spatial attention modules are proposed to capture the global spatial correlations between different regions at node and trend levels; 2) From a global temporal view, a temporal attention module is employed to capture and encode global temporal correlations between traffic at different times; 3) From a local spatial-temporal view, a dense convolution module is developed to further excavate the local spatial-temporal dependencies in cellular traffic. Consequently, a successful cellular traffic prediction strategy is constructed to fully explore the spatial-temporal characteristics from multiple views. The experimental results on a popular real-world cellular traffic dataset demonstrate that the MVSTGN achieves obvious improvements over baselines.

In this paper, we focus on a wireless-powered sensor network coordinated by a multi-antenna access point (AP). Each node can generate sensing information and report the latest information to the AP using the energy harvested from the AP’s signal beamforming. We aim to minimize the average age-of-information (AoI) by adapting the nodes’ scheduling and the transmission control strategies jointly. To reduce the transmission delay, an intelligent reflecting surface (IRS) is used to enhance the channel conditions by controlling the AP’s beamforming strategy and the IRS’s phase shifting matrix. Considering dynamic data arrivals at different sensing nodes, we propose a hierarchical deep reinforcement learning (DRL) framework for AoI minimization in two steps. The users’ transmission scheduling is firstly determined by the outer-loop DRL approach, e.g. the DQN or PPO algorithm, and then the inner-loop optimization is used to adapt either the uplink information transmission or downlink energy transfer to all nodes. A simple and efficient approximation is also proposed to reduce the inner-loop rum time overhead. Numerical results verify that the hierarchical learning framework outperforms typical baselines in terms of the average AoI and proportional fairness among different nodes.

With the accelerated popularization of 5G applications, accurate cellular traffic prediction is becoming increasingly important for efficient network management. Currently, the latest algorithms for cellular traffic prediction generally neglect extraction of the shallow features of cellular traffic and the prediction accuracy is hence limited. Therefore, we propose a global-local spatial-temporal transformer network (GLSTTN) that can fully excavate diverse spatial-temporal characteristics of cellular traffic for accurate cellular traffic prediction. Specifically, GLSTTN achieves this goal by constructing two modules: the global spatial-temporal module and the local spatial-temporal module. In the global spatial-temporal module, GLSTTN captures global correlations using stacked spatial-temporal blocks, where each block is composed of one spatial transformer and one temporal transformer. A skip connection is then used in each block to strengthen feature propagation. In the local spatial-temporal module, GLSTTN fully extracts the local spatial-temporal dependencies hidden in globally encoded features using densely connected convolutional neural networks. Extensive experiments demonstrate that GLSTTN achieves more accurate cellular traffic prediction than existing approaches on a real-world cellular traffic dataset.

The Third Generation Partnership Project has standardized cellular vehicle-to-everything (C-V2X) sidelink Mode 4 communication to support vehicle-to-vehicle safety applications. In Mode 4, the sensing-based semi-persistent scheduling (SPS) scheme allows vehicles to select radio resources autonomously. In particular, SPS has three steps to generate available resource lists (ARLs) for resource selection. However, the overlapping of ARLs is inevitable when radio resources are insufficient. In this case, randomly selecting from ARLs according to SPS is likely to make adjacent vehicles select the same resources, resulting in packet collisions. Unlike SPS, this paper proposes a multiagent deep reinforcement learning-based SPS (RL-SPS) algorithm to help vehicles select appropriate radio resources. As a consequence, packet collisions can be largely reduced. Furthermore, a centralized multi-head attention mechanism is adopted to improve the efficiency of the training process of RL-SPS. Simulation results demonstrate the reliability, scalability and robustness of the RL-SPS in a dynamic vehicular network.

Mobile crowdsensing (MCS) is a new sensing paradigm that relies on the crowd's sensing capabilities to aggregate data. Unlike traditional MCS systems, where sensing data are traded via a third-party platform, we propose a blockchain-based data trading framework to ensure the security of data transactions in the MCS system. In particular, the interactions between selling mobile users (SMUs) and buying mobile users (BMUs) are modeled as a Stackelberg game. Then, the optimal unit price and the amount of sensing time purchased from SMUs are solved by two smart contracts. Notably, the SMUs are compensated according to not only the amount of sensing time but also their historical reputation to encourage SMUs to contribute high-quality data. Furthermore, the blockchain technology guarantees that the reputations of each SMU are recorded in a traceable manner. Experimental results confirm that the proposed mechanism achieves near-optimal social welfare while protecting the security of data transaction.

A computation-intensive graph task comprises a set of tasks and corresponding data flows between adjacent tasks. In this paper, we consider a mobile edge computing (MEC) system based on vehicle-to-everything (V2X) communication in which each task initiator (TI) generates and offloads a set of correlated tasks to different task executors (TEs). We formulate the graph task assignment problem as a mixed-integer nonlinear programming problem (MINLP) to minimize the weighted sum of time-energy consumption (WETC). Due to the data-flow dependency and time-varying characteristics of the operating environment including channel gain, communication distance between TIs and TEs, available computing resources of TEs, traditional numerical optimization algorithms cannot solve such optimization problem efficiently, especially when the scale of the MEC system is quite large. To this end, we propose a graph task offloading mechanism named GT-DQN by integrating deep Q-Network (DQN) with breadth-first search technique. Firstly, DQN is trained to generate a near-optimal offloading strategy, through numerous interactions with the time-varying operating environment. Secondly, a breadth-first search algorithm is adopted to traverse the graph task, which can significantly reducing the computational complexity. Compared with existing algorithms, simulation results verify the superiority of GT-DQN.

Multi-access edge computing (MEC) and ultra-dense networking (UDN) are recognized as two promising paradigms for future mobile networks that can be utilized to improve the spectrum efficiency and the quality of computational experience (QoCE). In this paper, we study the task offloading problem in an MEC-enabled UDN architecture with the aim to minimize the task duration while satisfying the energy budget constraints. Due to the dynamics associated with the environment and parameter uncertainty, designing an optimal task offloading algorithm is highly challenging. Consequently, we propose an online task offloading algorithm based on a state-of-the-art deep reinforcement learning (DRL) technique: asynchronous advantage actor-critic (A3C). It is worthy of remark that the proposed method requires neither instantaneous channel state information (CSI) nor prior knowledge of the computational capabilities of the base stations. Simulations show that the our method is able to learn a good offloading policy to obtain a near-optimal task allocation while meeting energy budget constraints of mobile devices in UDN environment.

Long Range (LoRa) technology which provides low power consumption and long transmission range is regarded as one of the key technologies for industrial Internet of Things (IIoTs). In this article, we investigate an energy efficiency (EE) maximization problem through the joint allocation of spreading factors (SFs), channels (CHs) and transmit power. Then, the optimal solution is derived by decomposing the problem into two stages: (i) a deep Q-Network (DQN) is trained to generate the allocation of CH/SF based on the channel state information (CSI) collected from end-devices (EDs); (ii) given the CH/SF allocation, the optimal transmission power is then determined by solving a convex optimization problem. Simulation results demonstrate that our algorithm is superior to the random CH/SF and power selecting method, and achieves a near-optimal performance.

Digital twin is the innovation backbone of the smart manufacturing by delivering virtual representation of the real world. Aiming at constructing virtual representations of visual scenes, scene graph generation is a digital twin task that not only models objects but also infers their relationships. Existing works usually learn coarse global context when predicting relationships leading to excessive redundant information being considered. In this article, we first classify objects into different subgroups according to the degree of correlations with several hub objects. Then, we propose a multihub driven attention network (MHDANet) based on deep learning that drives the information to pass within the subgroups and forces objects to attend more to related objects. Consequently, MHDANet learns compact relation-aware features of visual scenes and predicts accurate and diverse relationships. Experimental results show that MHDANet achieves superb performance on scene graph generation on real-world datasets and especially alleviates the imbalance of predicted relationship categories.

We investigate how to provide Quality-of-Service (QoS) for diversified video flows. We design a fine-grained video traffic control mechanism that integrates traffic classification with path selection for video flows within the framework of SDN. For the design, we present a category-theoretic ontology log (olog) diagram model, which provides a novel perspective on the interdependency among various system components. For the video traffic classification, we first evaluate various machine learning classifiers in terms of their performance and then chose the most effective one to be the first module. For the path selection, we devise a multi-constrained QoS routing strategy by restructuring a state-of-the-art graph algorithm, combine it with the${k}$-shortest path algorithm, and deploy this strategy as another video traffic control module. We implemented a prototype of the proposed mechanism on the SDN emulator Mininet, and we evaluate its effectiveness using the performance results obtained.

Multi-access edge computing (MEC) and ultra-dense networking (UDN) are recognized as two promising paradigms for future mobile networks that can be utilized to improve the spectrum efficiency and the quality of computational experience (QoCE) . In this paper, we study the task offloading problem in an MEC-enabled UDN architecture with the aim to minimize the task duration while satisfying the energy budget constraints. Due to the dynamics associated with the environment and parameter uncertainty, designing an optimal task offloading algorithm is highly challenging. Consequently, we propose an online task offloading algorithm based on a state-of-the-art deep reinforcement learning (DRL) technique: asynchronous advantage actor-critic (A3C) . It is worthy of remark that the proposed method requires neither instantaneous channel state information (CSI) nor prior knowledge of the computational capabilities of the base stations. Simulations show that the our method is able to learn a good offloading policy to obtain a near-optimal task allocation while meeting energy budget constraints of mobile devices in the UDN environment.

Existing works of scene graph generation usually learn global context information leading to excessive redundant information being considered. As a matter of fact, most relationships only exist around several hub objects. To exploit this fact, we propose a novel Multiple Hub-driven Attention Graph Network (MHAGN) for scene graph generation. Specifically, we first classify objects into multiple subgraphs according to the degree of correlations with the hub objects. Then, we design Multiple Hub-driven Attention (MHA) to drive context information to be passed within multiple subgraphs separately, and force objects to attend more to associated objects in the sub-graph. Finally, MHAGN captures precise and diverse context information by combining MHAs from multiple subgraphs and generates compact relation-aware representations for objects. Experimental results on popular benchmarks show that the proposed MHAGN achieves better performance over baselines on several datasets, especially in terms of alleviating the imbalance of predicted relationship categories.

Ultrareliable and low-latency communication (URLLC) is a prerequisite for the successful implementation of the Internet of Controllable Things. In this article, we investigate the potential of deep reinforcement learning (DRL) for joint subcarrier-power allocation to achieve low latency and high reliability in a general form of device-to-device (D2D) networks, where each subcarrier can be allocated to multiple D2D pairs and each D2D pair is permitted to utilize multiple subcarriers. We first formulate the above problem as a Markov decision process and then propose a double deep $Q$ -network (DQN)-based resource allocation algorithm to learn the optimal policy in the absence of full instantaneous channel state information (CSI). Specifically, each D2D pair acts as a learning agent that adjusts its own subcarrier-power allocation strategy iteratively through interactions with the operating environment in a trial-and-error fashion. Simulation results demonstrate that the proposed algorithm achieves near-optimal performance in real time. It is worth mentioning that the proposed algorithm is especially suitable for cases where the environmental dynamics are not accurate and the CSI delay cannot be ignored.

Mobile crowdsensing (MCS) is an appealing sensing paradigm that leverages the sensing capabilities of smart devices and the inherent mobility of device owners to accomplish sensing tasks with the aim of constructing powerful industrial systems. Incentivizing mobile users (MUs) to participate in sensing activities and contribute high-quality data is of paramount importance to the success of MCS services. In this article, we formulate the competitive interactions between a sensing platform (SP) and MUs as a multistage Stackelberg game with the SP as the leader player and the MUs as the followers. Given the unit prices announced by MUs, the SP calculates the quantity of sensing time to purchase from each MU by solving a convex optimization problem. Then, each follower observes the trading records and iteratively adjusts their pricing strategy in a trial-and-error manner based on a multiagent deep reinforcement learning algorithm. Simulation results demonstrate the efficiency of the proposed method.

It is expected that a mixture of autonomous and manual vehicles will persist as a part of the intelligent transportation system (ITS) for many decades. Thus, addressing the safety issues arising from this mix of autonomous and manual vehicles before autonomous vehicles are entirely popularized is crucial. As the ITS system has increased in complexity, autonomous vehicles exhibit problems such as a low intention recognition rate and poor real-time performance when predicting the driving direction; these problems seriously affect the safety and comfort of mixed traffic systems. Therefore, the ability of autonomous vehicles to predict the driving direction in real time according to the surrounding traffic environment must be improved and researchers must work to create a more mature ITS. In this paper, we propose a deep learning-based traffic safety solution for a mixture of autonomous and manual vehicles in a 5G-enabled ITS. In this scheme, a driving trajectory dataset and a natural-driving dataset are employed as the network inputs to long-term memory networks in the 5G-enabled ITS: the probability matrix of each intention is calculated by the softmax function. Then, the final intention probability is obtained by fusing the mean rule in the decision layer. Experimental results show that the proposed scheme achieves intention recognition rates of 91.58% and 90.88% for left and right lane changes, respectively, effectively improving both accuracy and real-time intention recognition and improving the lane change problem in a mixed traffic environment.

Device-to-device communications underlaying cellular networks have been recognized as one of the key technologies for the fifth generation (5G) cellular system to improve the spectrum efficiency and system capacity. In this paper, we investigate the potential of deep reinforcement learning (DRL) for joint subcarrier assignment and power allocation in a general form of D2D networks, where a subcarrier can be assigned to multiple D2D pairs and each D2D pair is permitted to utilize multiple subcarriers. We first formulate the above problem as a Markov decision process, and then propose a double deep Q-network (DQN)-based subcarrier-power allocation algorithm to learn the optimal policy in the absence of full instantaneous channel state information (CSI). Specifically, each D2D pair acts as a learning agent that adjusts its own subcarrier-power allocation strategy iteratively through interactions with the operating environment in a trial-and-error fashion. Simulation results confirm that the proposed algorithm achieves near optimal performance in real time. It is worth mentioning that the proposed algorithm is especially suitable for the case where the environmental dynamics is not accurate and the CSI delay cannot be ignored.

Due to the bursty nature of Internet traffic, network service providers (NSPs) are forced to expand their network capacity in order to meet the ever-increasing peak-time traffic demand, which is however costly and inefficient. How to shift the traffic demand from peak time to off-peak time is a challenging task for NSPs. In this paper, we study the implementation of time-dependent pricing (TDP) for bandwidth slicing in software-defined cellular networks under information asymmetry and price discrimination. Congestion prices indicating real-time congestion levels of different links are used as a signal to motivate delay-tolerant users to defer their traffic demands. We formulate the joint pricing and bandwidth demand optimization problem as a two-stage Stackelberg leader-follower game. Then, we investigate how to derive the optimal solutions under the scenarios of both complete and incomplete information. We also extend the results from the simplified case of a single congested link to the more complicated case of multiple congested links, where price discrimination is employed to dynamically adjust the price of each congested link in accordance with its real-time congestion level. Simulation results demonstrate that the proposed pricing scheme achieves superior performance in increasing the NSP's revenue and reducing the peak-to-average traffic ratio (PATR).

V2X task offloading, which moves computational tasks to nearby edge nodes (ENs) with available computing resources, has attracted attention. The task is offloaded to ENs that satisfies the task delay constraints, including the transmission delay and the calculation delay when the task is offloaded. However, since many existing methods assume a static environment in which nodes are fixed, communication may be disconnected in the dynamic environment in which nodes move. On the other hand, the task offloading methods in a dynamic environment, they do not consider the case where the node turning behavior, since they assume only the case where the nodes move in a straight line. Therefore, in order to establish appropriate ENs selection method in dynamic environment, we propose a V2V task offloading method considering automobile behavior including both going straight line and turning in urban area. And the effectiveness of the proposed method is shown in comparison with the existing method by simulation.

In this paper, we emphasize on energy-efficient resource allocation for the energy harvesting based cognitive machine-to-machine (EH-CM2M) communication. We consider how to maximize the energy efficiency of M2M transmitters (M2M-TXs) via the joint optimization of channel selection, peer discovery, power control, and time allocation. We propose a two-stage three-dimensional matching algorithm. In the first stage, M2M-TXs, M2M receivers (M2M-RXs) and resource blocks (RBs) are temporally matched together, and then the joint power control and time allocation problem is solved by combining alternating optimization (AO), nonlinear fractional programming, and linear programming to construct the preference lists. In the second stage, the joint channel selection and peer discovery problem is solved by the proposed pricing-based matching algorithm based on the established preference lists. Simulation results confirm that the proposed algorithm can approach the optimal performance with a low complexity.

Adaptive bitrate control (ABR) is an important technique for video streaming. This technique selects the video quality adaptively according to various network conditions, to ensure the quality of experience (QoE) for users. In the previous works, the ABR methods are mainly tested in single user environment. In this paper, an emulation testbed is constructed for QoE performance evaluation in multi-user networks. The state-of-the-art ABR methods are incorporated into the proposed environment. Emulation experiments are carried out to evaluate the performance of the methods. Preliminary results show that in FESTIVE, which has the least QoE variation in the six-user experiments, the user QoE of the worst case is only 27.5% of that of best case, demonstrating that the state-of-the-art ABR methods are not effective enough to optimize the QoE for all users under multi-user condition. Future design of the ABR method should take factors such as the fairness and resource allocation into consideration.

Multi-Access Edge Computing (MEC) is an emerging technology that leverages computing, storage and network resources deployed at the proximity of users to offload terminal from computational- and delay-sensitive tasks. Various existing facilities including mobile devices with idle resources, vehicles, and MEC servers deployed at base stations or road side units, could act as edges in the network. Since offloading tasks incurs extra transmission energy consumption and transmission latency, two key questions to be addressed in MEC deployments are: (i) offload the workload to the edge or compute it in terminals? (ii) which edge, among the available ones, should the task be offloaded to? Hence, we propose a matching theory based task assignment mechanism which takes into account the devices' and MEC servers' computation capabilities, wireless channel conditions, and delay constraints. The main goal of our task assignment mechanism is to reduce overall energy consumption, while satisfying task owners' heterogeneous delay requirements and supporting good scalability. Simulations are conducted to evaluate the efficiency of our proposed mechanism.

Video caching is one of the most important research issues in Content-Centric Network (CCN) and greatly affects its overall performance. The computational complexity of state-of-the-art optimal caching schemes is high, due to the arbitrary network topologies. In this paper, the popularity-aware video caching in topology-known CCN is studied. The complex arbitrary network typology is mapped into a virtual cascade network topology and a caching scheme is designed in accordance with the transformed virtual network rather than the original network. This scheme is proved optimal, and is with polynomial computational complexity. Simulations are conducted and the results show that the proposed scheme outperforms the existing schemes.

Since errors are very common in the scientific and engineering data and/or evaluation processes. A reliable computation outsourcing scheme should not introduce extra error gains. However, we observed the existing matrix computation outsourcing schemes could not reconcile the reliability and security. In this paper, we introduce a new fully homomorphic matrix disguising scheme, based on sparse unitary disguising matrices, which take into account reliability and security. Our scheme can be regarded as a variant of matrix fully homomorphic applications in Symmetric key cryptosystem also.

In this paper, we propose a time-dependent pricing (TDP) scheme for bandwidth consumption scheduling of multimedia streaming applications. By decoupling the network control functions from data delivery, software defined network (SDN) enables multimedia streaming users to negotiate their QoS parameters in a on-demand basis. Our key idea is to employ TDP as an incentive mechanism in SDN to motivate users to shift their delay-tolerant traffic demand, and free up resources for delay-sensitive applications in peak-time. Then, a Stackelberg game is formulated to analyze the interactions between the ISP and users. Next, we proposed a greedy algorithm to obtain the optimal congestion price and bandwidth slicing in each time slot based on real-time traffic load as well as users' preferences for delay. Finally, simulation results confirm that the proposed method can significantly flatten out traffic fluctuation.

With the increasing popularity of unmanned aerial vehicles (UAVs), it is foreseen that they will play an important role in broadening the horizon of mobile crowd sensing (MCS). Specifically, UAV-aided MCS allows autonomous data collection anytime and anywhere due to the capability of fast deployment and controllable mobility. However, the on-board battery capacity of UAVs imposes a limitation on their endurance capability and performance. In this paper, we consider the fixed-wing UAV-aided MCS system and investigate the corresponding joint route planning and task assignment problem from an energy efficiency perspective. The formulated joint optimization problem is transformed into a two-sided two-stage matching problem, in which the route planning problem is solved in the first stage based on either dynamic programming or genetic algorithms, and the task assignment problem is addressed in the second stage by exploring the Gale-Shapley algorithm. We provide a comprehensive theoretical analysis, and elaborate the procedures of practical implementation. Numerical results demonstrate that significant performance improvement can be achieved by the proposed scheme.

Software-defined networking (SDN) enables network subscribers to negotiate QoS parameters in a on-demand basis. On the other hand, peak-time congestion accompanying the fast-growing traffic in recent years forces Internet service provider (ISP) to put forward a new pricing scheme by taking into account when a user uses Internet in addition to how much a user uses Internet. In this paper, we study the payoff optimization problem of ISP and network subscribers in SDN-enabled networks. A self-interested network subscriber always tries to obtain network resources as much as possible even if the network is congested; on the other hand, rational ISP tends to charge a higher price without providing subscribers guaranteed Quality of Service (QoS). A Stackelberg game is hence constructed to analyze the competitive interactions between ISP and home network subscribers. Specifically, ISP decides its pricing strategy for each time slot by solving a payoff optimization problem. Given the pricing strategy, network subscribers then decide the bandwidth to be reserved in a on-demand basis aiming to optimize their own payoff as well. We analyze the Nash equilibrium solution of the game. Simulation results confirm that the proposed pricing scheme can largely improve the payoff of network subscribers and ISP, compared to the usage-based pricing (UBP) scheme. Furthermore, the portion of surplus obtained by ISP increases with the increase of traffic load.

This paper proposes a smart grid policy simulation platform, which takes all information affecting the interests of the stakeholders into consideration, to set policy and conduct empirical analysis. The platform can help policy making departments to understand the policy implementation effect (e.g., environmental benefits) more intuitively by using visual analysis and reporting tools, so as to find a best policy combination. The platform uses Python to realize the united simulation of Matlab and Gridview to calculate systematical carbon emissions reduction. Thus, the proposed platform can predict whether the goals of energy conservation and emissions reduction could be achieved. The paper takes IEEE RTS-25 node system as a demonstration case to calculate carbon emission reduction and get policy combination of different targets taking into account the subsidies for wind power, carbon tax, and electricity price of thermal power.

The past few years have witnessed a huge acceleration in global Internet traffic. Users' demand for contents is also rising accordingly. Therefore, content providers (CPs) that provide contents for users get high revenue from the traffic growth. There are generally two ways for CPs to get revenue: (i) charge users for the contents they view or download; (ii) get revenue from advertisers. On the other hand, Internet service providers (ISPs) are investing in network infrastructure to provide better quality of service (QoS), but they do not benefit directly from the content traffic. One option for ISPs to compensate their investment cost is sharing CPs' revenue by side payment from CPs to ISPs. Then ISPs will be motivated to keep on investing in developing new network technology and enlarging the capacity to improve QoS. However, it is important to evaluate how each player is affected by this kind of side payment. Our previous work has studied this problem by assuming that CPs charged users for the contents they view or download, in this paper it is considered that CP does not directly charge end users, but charges advertisers for revenue. Stackelberg game is utilized to study the interactions among ISP, CP, end users and advertisers. A unique Nash equilibrium is established and numerical analysis has validated our theoretic results. It shows that side payment from CP to ISP impairs the CP's investment of contents, and ISP can benefit from charging CP, while CP's payoff is impaired.

In this paper, we consider the problem of spectrum sharing in a Cognitive Radio Network (CRN) with spectrum holder, two secondary operators and secondary users (SUs). In the system model under consideration, the spectrum allocated to the two secondary operators can be shared by SUs, which means that secondary operators buy spectrum from spectrum holder and then sell spectrum access service to SUs. We model the relationship between secondary operators and SUs as a two-stage stackelberg game, where secondary operators make spectrum channel quality and price decisions in the first stage, and then the SUs make their spectrum demands decisions. The backward induction method is employed to solve the stackelberg game. Numerical results are performed to evaluate our analysis.

Nowadays, mobile network operators (MNOs) are trying to deploy wireless local area network (LAN) to offload mobile data from their cellular networks to complementary wireless LAN for congestion relief and cost savings. However, these network-centric methods do not take into consideration mobile user's (MU's) interests of monetary cost, energy consumption, and applications' deadlines. How the MU decides whether to offload their traffic to a complementary wireless LAN is non-trivial and important issue. Previous studies assume that MNO adopts usage-based pricing for mobile data, which only cares about how much a MU consumes data but not when a MU consumes data. In this paper, we study the MU's policy to minimize his monetary cost and energy consumption under time-dependent pricing (TDP). We formulate MU's wireless LAN offloading problem as a finite-horizon discrete-time Markov decision process (MDP) and establish an optimal policy by a dynamic programming based algorithm. Extensive simulations are conducted to validate our proposed offloading algorithm.

In order to improve the quality of service (QoS) for mobile users (MUs) and save investment cost for deploying new cellular base station, mobile network operators (MNOs) are deploying wireless local area network (LAN) access points (APs) to offload MU's traffic from cellular network to wireless LAN. However, offloading too much traffic from cellular network may impair MNO's profit since the cellular network price is higher than that of wireless LAN, whose price is low or even zero. Therefore, how to deploy wireless LAN APs to offload traffic without impairing MNO's profit is a critical problem for MNOs. As far as the authors understand, existing studies about deployment of wireless LAN APs do not consider MNO's profit and are usually in a heuristic manner. In this paper, we study the location-based advertising (LBA) leveraged wireless LAN deployment, where MNO may also collect revenue by selling LBA service in different locations to advertisers. We formulate MNO's profit maximization problem by considering different MU's demand in different locations, wireless LAN price for MUs, and revenue from LBA service. Extensive simulations are conducted to validate our analytical results.

The ongoing evolution of personal mobile devices capabilities and wireless technologies result in a huge growth of traffic on mobile networks (3G/4G). One of the most promising approaches to handle this data crisis is to offload the fast growing traffic onto femtocell networks. Since both the 3G/4G macrocell and femtocells operate on the same licensed spectrum, the cross-tier interference should be well managed. In this paper, we propose a utility-based transmission power allocation policy for the uplink transmission in femtocell networks. Our main motivation is to design the transmission power allocation policy aiming at optimizing the joint utility of femtocell users (FUs) subject to a interference temperature constraint at the macrocell base station (MBS) side. We provide a novel floating-ceiling water-filling (FCWF) algorithm with little computational overhead to obtain the optimal solution for the joint utility optimization problem. Numerical results confirm that the joint utility and average SINR can be significant improved with the proposed method.

Content Centric Network (CCN) has become a heated research topic recently, as it is proposed as an alternative of the future network. The routers in CCN have the caching abilities and the caching strategies affect the system performance greatly. Each content in CCN is associated with a popularity, which is determined by the corresponding requested times. Popularity-aware caching scheme caches the popular content close to users and can lead to better caching performance in terms of smaller average transmission hops traveled. Content popularity significantly affects the overall system performance, and the content size is not considered during the content level popularity (i.e. item popularity) calculation. In this paper, we study the impact of the item popularity and chunk popularity in CCN, where the chunk popularity is the normalized item popularity considering the content size. Extensive simulations are conducted and the simulation results show the advantages and disadvantages of each scheme. A new popularity calculation method is proposed to perform the tradeoff between the item popularity and chunk popularity.

The ongoing enhancement of mobile devices capabilities and advanced technologies has largely contributed to the huge increase of wireless traffic demand. Mobile network operators around the world are struggling to manage this new era of higher generation networks by finding new means and strategies to implement in order to meet the explosion in traffic demand. One of the current and most efficient strategies to handle this issue is the femtocell technology, which now has drawn the interest of main concerned actors as for manufacturers, operators, and researchers. In this paper, our main motivation was to address the macro-femto heterogeneous networks deployment issue from the economic side. Thus, we aimed to propose a pricing strategy for macro-femto networks with a user centric vision where users would have the choice to access one of the networks based on the proposed service price when he/she is accessing it. Based on our strategy, a wireless service provider's (WSP's) revenue optimization is then performed and evaluated in order to show the efficiency of our proposed pricing scheme.

With rapid increases in demand for mobile data, mobile network operators are trying to expand wireless network capacity by deploying WiFi hotspots to offload their mobile traffic. However, these network-centric methods usually do not fulfill interests of mobile users (MUs). MUs consider many problems to decide whether to offload their traffic to a complementary WiFi network. In this paper, we study the WiFi offloading problem from MU's perspective by considering delay-tolerance of traffic, monetary cost, energy consumption as well as the availability of MU's mobility pattern. We first formulate the WiFi offloading problem as a finite-horizon discrete-time Markov decision process (FDTMDP) with known MU's mobility pattern and propose a dynamic programming based offloading algorithm. Since MU's mobility pattern may not be known in advance, we then propose a reinforcement learning based offloading algorithm, which can work well with unknown MU's mobility pattern. Extensive simulations are conducted to validate our proposed offloading algorithms.

Network traffic load usually differs significantly at different times of a day due to users' different time preference. Network congestion may happen in traffic peak times. In order to prevent this from happening, network service providers (NSPs) can either over-provision capacity for demand at peak times of the day, or use dynamic time-dependent pricing (TDP) scheme to reduce the demand at traffic peak times. Since over-provisioning network capacity is costly, many researchers have proposed TDP schemes to control congestion as well as to improve the revenue of NSPs. To the best of our knowledge, all these studies consider only the monopoly NSP case. In our previous work, the duopoly and oligopoly NSP cases have been studied. NSPs try to maximize their overall revenue by setting time-dependent prices, while users choose NSPs by considering their own time preference, congestion statuses in the networks and the prices set by the NSPs. One assumption that has been made is that Quality of Service (QoS) function of each NSP is linear, which means that the level of QoS degradation is proportional to the number of users in the network. However, in reality, the level of QoS may degrade rapidly after a certain point, which is not reflected through linear QoS functions. Therefore, concave QoS function is a better choice. In this paper, the case of concave QoS function is considered. TDP is evaluated under different QoS functions. The results shows that TDP is also effective under concave QoS functions.

Pricing cloud resources plays an important role in leading to the success of cloud computing. Cloud services are priced at different levels in infrastructure-as-a-service (IaaS) cloud market. For example, Amazon EC2 offers its cloud resources with three pricing schemes, the subscription model, pay-as-you-go model and spot pricing model. With more and more IaaS cloud service providers (CSPs) beginning to provide cloud services, they form a competitive market to compete for cloud users. Therefore, how to set optimal prices in order to maximize their revenue in a competitive IaaS cloud computing market while at the same time meeting the cloud users' demand satisfaction is a problem that CSPs should consider. Towards this end, in this paper, we study subscription pricing competition in a duopoly IaaS cloud computing market. First, we analyze whether or not the cloud users choose to use cloud service. Then, we present a game theoretic analysis of a cloud market with two CSPs competing non-cooperatively for cloud users.

In this paper, we consider a price-based connection admission control (CAC) for wireless random access networks. In particular, a network operator determines sequential prices to dynamically maintain the traffic admitted into the network below the channel capacity. The CAC tries to ensure quality of service (QoS) guarantees to users and hence maximize the overall revenue. We find that the revenue maximization problem over all sequential prices is NP-hard. Therefore, a greedy algorithm is employed for obtaining a simple, easy-to-implement solution.