The World Wide Web thrives on intelligent services that rely on accurate time series classification, which has recently witnessed significant progress driven by advances in deep learning. However, existing studies face challenges in domain incremental learning. In this paper, we propose a lightweight and robust dual-causal disentanglement framework (DualCD) to enhance the robustness of models under domain incremental scenarios, which can be seamlessly integrated into time series classification models. Specifically, DualCD first introduces a temporal feature disentanglement module to capture class-causal features and spurious features. The causal features can offer sufficient predictive power to support the classifier in domain incremental learning settings. To accurately capture these causal features, we further design a dual-causal intervention mechanism to eliminate the influence of both intra-class and inter-class confounding features. This mechanism constructs variant samples by combining the current class's causal features with intra-class spurious features and with causal features from other classes. The causal intervention loss encourages the model to accurately predict the labels of these variant samples based solely on the causal features. Extensive experiments on multiple datasets and models demonstrate that DualCD effectively improves performance in domain incremental scenarios. We summarize our rich experiments into a comprehensive benchmark to facilitate research in domain incremental time series classification.

Although Transformers excel in natural language processing, their extension to time series forecasting remains challenging due to insufficient consideration of the differences between textual and temporal modalities. In this paper, we develop a novel Transformer architecture designed for time series data, aiming to maximize its representational capacity. We identify two key but often overlooked characteristics of time series: (1) unidirectional influence from the past to the future, and (2) the phenomenon of decaying influence over time. These characteristics are introduced to enhance the attention mechanism of Transformers. We propose TimeFormer, whose core innovation is a self-attention mechanism with two modulation terms (MoSA), designed to capture these temporal priors of time series under the constraints of the Hawkes process and causal masking. Additionally, TimeFormer introduces a framework based on multi-scale and subsequence analysis to capture semantic dependencies at different temporal scales, enriching the temporal dependencies. Extensive experiments conducted on multiple real-world datasets show that TimeFormer significantly outperforms state-of-the-art methods, achieving up to a 7.45% reduction in MSE compared to the best baseline and setting new benchmarks on 94.04% of evaluation metrics. Moreover, we demonstrate that the MoSA mechanism can be broadly applied to enhance the performance of other Transformer-based models.

Object categories are typically organized into a multi-granularity taxonomic hierarchy. When classifying categories at different hierarchy levels, traditional uni-modal approaches focus primarily on image features, revealing limitations in complex scenarios. Recent studies integrating Vision-Language Models (VLMs) with class hierarchies have shown promise, yet they fall short of fully exploiting the hierarchical relationships. These efforts are constrained by their inability to perform effectively across varied granularity of categories. To tackle this issue, we propose a novel framework (HGCLIP) that effectively combines CLIP with a deeper exploitation of the Hierarchical class structure via Graph representation learning. We explore constructing the class hierarchy into a graph, with its nodes representing the textual or image features of each category. After passing through a graph encoder, the textual features incorporate hierarchical structure information, while the image features emphasize class-aware features derived from prototypes through the attention mechanism. Our approach demonstrates significant improvements on 11 diverse visual recognition benchmarks. Our codes are fully available at https://github.com/richard-peng-xia/HGCLIP.

Currently, most spiking neural networks (SNNs) still mimic the chain-like hierarchical architecture in traditional artificial neural networks (ANNs). This method significantly differs from random connections between neurons found in biological brains, limiting the ability to model the evolving mechanisms of neural pathways in biological neural systems, particularly in terms of dynamic depth-scalability and adaptive path-plasticity. This paper develops a new modeling paradigm for SNNs with random graph architecture (RGA), termed Cognition-aware SNN (CogniSNN). Furthermore, we model the depth-scalability and path-plasticity in CogniSNN by introducing a modified spiking residual neural node (ResNode) to counteract network degradation in deeper graph pathways, as well as a critical path-based algorithm that enables CogniSNN to perform path reusability on new tasks leveraging the features of the data and the RGA learned in old tasks. Experiments show that the performance of CogniSNN with redesigned ResNode is comparable, even superior, to current state-of-the-art SNNs on neuromorphic datasets. The critical path-based approach effectively achieves path reuse capability while maintaining expected performance in learning new tasks that are similar to or distinct from the old ones. This study showcases the potential of RGA-based SNNs and paves a new path for modeling the fusion of computational neuroscience and deep intelligent agents. The code is available at github.com/Yongsheng124/CogniSNN.

Stock trend prediction involves forecasting the future price movements by analyzing historical data and various market indicators. With the advancement of machine learning, graph neural networks (GNNs) have been extensively employed in stock prediction due to their powerful capability to capture spatiotemporal dependencies of stocks. However, despite the efforts of various GNN stock predictors to enhance predictive performance, the improvements remain limited, as they focus solely on analyzing historical spatiotemporal dependencies, overlooking the correlation between historical and future patterns. In this study, we propose a novel distillation-based future-aware GNN framework (DishFT-GNN) for stock trend prediction. Specifically, DishFT-GNN trains a teacher model and a student model, iteratively. The teacher model learns to capture the correlation between distribution shifts of historical and future data, which is then utilized as intermediate supervision to guide the student model to learn future-aware spatiotemporal embeddings for accurate prediction. Through extensive experiments on two real-world datasets, we verify the state-of-the-art performance of DishFT-GNN.

The Spiking Neural Network (SNN) has drawn increasing attention for its energy-efficient, event-driven processing and biological plausibility. To train SNNs via backpropagation, surrogate gradients are used to approximate the non-differentiable spike function, but they only maintain nonzero derivatives within a narrow range of membrane potentials near the firing threshold—referred to as the surrogate gradient support width gamma. We identify a major challenge, termed the dilemma of gamma: a relatively large gamma leads to overactivation, characterized by excessive neuron firing, which in turn increases energy consumption, whereas a small gamma causes vanishing gradients and weakens temporal dependencies. To address this, we propose a temporal Inhibitory Leaky Integrate-and-Fire (ILIF) neuron model, inspired by biological inhibitory mechanisms. This model incorporates interconnected inhibitory units for membrane potential and current, effectively mitigating overactivation while preserving gradient propagation. Theoretical analysis demonstrates ILIF’s effectiveness in overcoming the gamma dilemma, and extensive experiments on multiple datasets show that ILIF improves energy efficiency by reducing firing rates, stabilizes training, and enhances accuracy. The code is available at github.com/kaisun1/ILIF.

Real-world time series typically exhibit complex temporal variations, making the time series classification task notably challenging. Recent advancements have demonstrated the potential of multi-scale analysis approaches, which provide an effective solution for capturing these complex temporal patterns. However, existing multi-scale analysis-based time series prediction methods fail to eliminate redundant scale-shared features across multi-scale time series, resulting in the model over- or under-focusing on scale-shared features. To address this issue, we propose a novel end-to-end Disentangled Multi-Scale framework for Time Series classification (DisMS-TS). The core idea of DisMS-TS is to eliminate redundant shared features in multi-scale time series, thereby improving prediction performance. Specifically, we propose a temporal disentanglement module to capture scale-shared and scale-specific temporal representations, respectively. Subsequently, to effectively learn both scale-shared and scale-specific temporal representations, we introduce two regularization terms that ensure the consistency of scale-shared representations and the disparity of scale-specific representations across all temporal scales. Extensive experiments conducted on multiple datasets validate the superiority of DisMS-TS over its competitive baselines, with the accuracy improvement up to 9.71%.

The morphological diversity, referring to the variations in traffic network topologies defined in this paper, often emerges and brings difficulties in successfully transferring a pre-trained prediction model from one traffic network to another. Moreover, most existing research primarily assumes that traffic data in source and target networks follow independent and identically distributed (i.i.d.) patterns, which is usually not consistent with real-world situations, particularly when considering morphological diversity. For this inconsistency, many efforts have been made, but they mainly concentrate on temporal aspects, which significantly differ from traffic prediction due to spatial and temporal correlations among road segments, influenced by variations in road topology and traffic behavior. This paper introduces a causality-based spatiotemporal out-of-distribution (OOD) generalization method, which is adaptable to most GNNs for diverse, large-scale, dynamic traffic systems with zero-shot. Furthermore, to enhance the generalization and adaptability of the proposed method, we introduce graph matching and equal-sized graph partitioning to alleviate spatial shift between the source and target traffic networks, reduce and align the scale of the networks. Experiments carried out on traffic flow datasets demonstrate that our method significantly improves the performance of various GNN-based traffic predictors in the situation of morphological diversity, achieving a maximum reduction in MAE of 33.08%. Compared to other OOD-driven baselines, our approach also shows a notable improvement, with up to a 40.58% decrease in MAE.

Covariance matrix estimation is an important problem in statistics, with wide applications in finance, neuroscience, meteorology, oceanography, and other fields. However, when the data are high-dimensional and constantly generated and updated in a streaming fashion, the covariance matrix estimation faces huge challenges, including the curse of dimensionality and limited memory space. The existing methods either assume sparsity, ignoring any possible common factor among the variables, or obtain poor performance in recovering the covariance matrix directly from sketched data. To address these issues, we propose a novel method - KEEF: Knowledge-based Time and Memory Efficient Covariance Estimator in Factor Model and its extended variation. Our method leverages historical data to train a knowledge-based sketch matrix, which is used to accelerate the factor analysis of streaming data and directly estimates the covariance matrix from the sketched data. We provide theoretical guarantees, showing the advantages of our method in terms of time and space complexity, as well as accuracy. We conduct extensive experiments on synthetic and real-world data, comparing KEEF with several state-of-the-art methods, demonstrating the superior performance of our method.

Precision matrix estimation is a ubiquitous task featuring numerous applications such as rare disease diagnosis and neural connectivity exploration. However, this task becomes challenging in small sample settings, where the number of samples is significantly less than the number of dimensions, leading to unreliable estimates. Previous approaches either fail to perform well in small sample settings or suffer from inefficient estimation processes, even when incorporating meta-learning techniques. To this end, we propose a novel approach FasMe for Fast and Sample-efficient Meta Precision Matrix Learning, which first extracts meta-knowledge through a multi-task learning diagram. Then, meta-knowledge constraints are applied using a maximum determinant matrix completion algorithm for the novel task. As a result, we reduce the sample size requirements to $O(\log p/K)$ per meta-training task and $O(\log\vert \mathcal{G}\vert)$ for the meta-testing task. Moreover, the hereby proposed model only needs $O(p \log\epsilon^{-1})$ time and $O(p)$ memory for converging to an $\epsilon$-accurate solution. On multiple synthetic and biomedical datasets, FasMe is at least ten times faster than the four baselines while promoting prediction accuracy in small sample settings.

With the proliferation of cellular vehicle-to-everything (C-V2X), connected and automated vehicles (CAVs) are gradually being commercialized. CAVs can interact with road infrastructure and human-driven vehicles (HDVs) to acquire relevant traffic information, thereby altering the characteristics of the traditional traffic flow. The emergence of CAVs is widely believed to bestow benefits to the traffic system in terms of safety, efficiency, and energy consumption. Nevertheless, as with most phenomena, there are two sides to the coin. Further exploration is necessary to determine whether the emergence of CAVs will trigger adverse effects and the underlying factors that may induce adverse effects. To be specific, this article first delves into how selfish driving behaviors (egoism CAV control strategy) can have an unfavorable impact on the performance of the traffic systems, thereby lowering the traffic efficiency. Subsequently, we develop an unselfish (altruism) CAV control strategy that aims to achieve the global optimization and improve the overall road operational capacity. Based on the simulation results obtained at different inflow and outflow rates on highway, it is evident that egoism driving behavior leads to a 11.55% decrease in average speed performance as compared to the noncontrol strategy, while altruism driving behavior results in a 20.14% improvement. Furthermore, we compare the proposed strategy with the current road infrastructure control, which only improves the average speed performance by 11.6%. This indicates that controlling CAVs has the potential to replace the deployment of the traditional road infrastructure, thereby optimizing the social and economic benefits. This article can provide insightful guidance for the future policy formulation in the transportation authorities, wherein the emergence of CAVs needs to be effectively regulated based on the altruism, thus fostering the establishment and development of a safe and efficient mixed traffic ecosystem.

With the proliferation of connected and automated vehicles (CAVs), densely distributed edge computing nodes have emerged on roadways. Consequently, leveraging CAVs as a mobile computing platform can integrate idle vehicle resources to provide computational services for ubiquitous Internet of Things (IoT) devices. Numerous studies have investigated task offloading strategy in the systems with full CAVs penetration. It is expected that the coexistence of CAVs and human-driven vehicles (HDVs) in mixed traffic systems will continue for a considerable period. However, due to the impact of HDVs on communication performance, the task offloading model designed for the systems with full CAVs penetration are no longer applicable in mixed traffic systems. We explore task offloading schemes using CAVs as a mobile computing platform in mixed traffic systems to address this issue. Specifically, we first model the communication model in mixed traffic systems, taking into account the influence of HDVs on link interference, the alteration of path loss due to the impact of HDVs on routing, and the additional sensing tasks arising from the inability of HDVs and CAVs to communicate. Subsequently, considering that delay and energy consumption are crucial factors affecting the performance of CAVs as a mobile computing platform, we formulate the task offloading scheme as an optimization problem. Additionally, we employ a distributed offloading based on deep learning (DODL) algorithm to obtain approximately optimal offloading decisions. Simulation results demonstrate the effectiveness of the proposed model in mixed traffic systems. By employing the DODL algorithm, the CAVs as a mobile computing platform can achieve enhanced performance in terms of convergence, thereby advancing the development of autonomous driving in mixed traffic systems.

Connected and Automated Vehicles (CAVs) are characterized by diverse communication attributes, embodying the trajectory of future automotive progress. Meanwhile, the transportation system will be in a mixed stage of CAVs and Human-Driven Vehicles (HDVs) for a long time. The study of communication capacity and strategies for mixed traffic systems is of great significance for the popularization of CAVs and the deployment of communication infrastructures. However, current research mainly focuses on the communication capacity analysis in the scenario with full penetration of CAVs, while the influence caused by HDVs on Vehicle-to-Vehicle (V2V) communications and the capacity analysis of connected vehicles in mixed traffic systems need further understanding. To address this issue, this paper considers the shadow fading caused by HDVs on wireless communication links and analyzes the communication capacity in mixed traffic systems. Specifically, we first synthesize the V2V and Vehicle-to-Infrastructure (V2I) communication modes to propose an analytical framework for vehicular network communication capacity in mixed traffic. Then, a predictive communication strategy is also provided that caches the required content at infrastructure in advance according to predicted vehicle trajectories to improve the capacity of vehicular networks in mixed traffic. Furthermore, the derived capacity analysis theorems reveal the communication capacity of mixed traffic is closely related to the CAV penetration rate, the vehicle arrival rate, and the infrastructure deployment interval. Simulation results prove the effectiveness of the proposed framework, and the proposed predictive communication strategy can increase the mixed traffic communication capacity compared to existing communication strategies. The theoretical results herein can guide the implementation of vehicular network applications and the design of communication strategies in mixed traffic systems.

As the number of connected and autonomous vehicles (CAVs) on road networks continues to increase, mixed transportation scenarios where CAVs and human-driven vehicles (HDVs) coexist are becoming more common. Establishing dedicated lanes (DLs) for CAVs is crucial for managing mixed traffic and improving road capacity. In this paper, we provide a theoretical analysis of the relationship between the market penetration rate (MPR) of CAVs and road capacity in both single-lane scenarios and multiple-lane scenarios with DLs. We derive a critical MPR for CAVs, at which they can be seamlessly accommodated within the DLs. Our numerical results show that CAVs should be prioritized to enter DLs first to optimize road capacity in mixed traffic. We also derive and validate the road capacity in multiple-lane scenarios and provide an optimal strategy for setting up DLs under varying MPRs to maximize road capacity. Overall, our study provides valuable insights into the significance of DLs for CAVs in mixed traffic and offers guidance on their implementation to improve road capacity.

The application of deep learning to nursing procedure activity understanding has the potential to greatly enhance the quality and safety of nurse-patient interactions. By utilizing the technique, we can facilitate training and education, improve quality control, and enable operational compliance monitoring. However, the development of automatic recognition systems in this field is currently hindered by the scarcity of appropriately labeled datasets. The existing video datasets pose several limitations: 1) these datasets are small-scale in size to support comprehensive investigations of nursing activity; 2) they primarily focus on single procedures, lacking expert-level annotations for various nursing procedures and action steps; and 3) they lack temporally localized annotations, which prevents the effective localization of targeted actions within longer video sequences. To mitigate these limitations, we propose NurViD, a large video dataset with expert-level annotation for nursing procedure activity understanding. NurViD consists of over 1.5k videos totaling 144 hours, making it approximately four times longer than the existing largest nursing activity datasets. Notably, it encompasses 51 distinct nursing procedures and 177 action steps, providing a much more comprehensive coverage compared to existing datasets that primarily focus on limited procedures. To evaluate the efficacy of current deep learning methods on nursing activity understanding, we establish three benchmarks on NurViD: procedure recognition on untrimmed videos, procedure and action recognition on trimmed videos, and action detection. Our benchmark and code will be available at https://github.com/minghu0830/NurViD-benchmark.

With the development of autonomous driving technology, it is foreseeable that connected and automated vehicles (CAVs) will be fully popularized in people’s lives. During this process, transportation systems are expected to evolve into the era of mixed autonomy, where CAVs and human-driven vehicles (HDVs) coexist in road networks and share available road resources. To materialize the much-anticipated potential of CAVs, a thorough understanding of CAVs’ effects on transportation systems is indispensable. On the one hand, attributing to advanced sensing, communication, and computation capabilities, CAVs provide opportunities to enhance mixed traffic safety, improve energy savings and suppress shockwave spread. On the other hand, due to advantages in large-scale information and cloud-computing resources, CAVs have the ability to occupy more road resources compared with HDVs, resulting in a reduction in the travel efficiency of HDVs, and even of the entire transportation systems. In this article, by clarifying the key differences between HDVs and CAVs, we comprehensively review the potential impacts of CAVs when they are appearing on road networks coexisting with HDVs. It can be regarded as the first-of-its-kind paper that systematically overviews the impacts of CAVs in the era of mixed autonomy on both positive and negative emotions. Specifically, the main focuses of this article are: 1) what are the key differences between CAVs and HDVs? 2) what are the positive impacts of CAVs’ appearance on mixed traffic systems? 3) will the introduction of CAVs cause some negative effects simultaneously? and 4) what kinds of strategies should be employed to relieve these negative effects? Hopefully, this article can not only call for an objective attitude toward the introduction of CAVs, but also provide foresighted advice to address possible challenges during the popularization of CAVs, so as to create a cooperative, safe, and efficient mixed traffic ecosystem.

In recent years, deep learning-based dehazing models have presented a momentum of dramatic growth. Unfortunately, most deep learning-based approaches heavily rely on synthetically hazed images for model training, which makes these methods brittle to restore hazy images taken from real-world scenes, due to the sample distribution discrepancy between synthetic and realistic images. Although some attempts have been made to overcome this difficulty by augmenting image spatial features with spectral features, the power of the spectral features still remains underutilized. In this paper, we propose the Spectral Dual-Channel Encoding (SDCE) framework for high-quality image dehazing, by unleashing the power of spectral feature encoding. We argue that hazes impose more adverse impacts on high-frequency image features (e.g., outlines and textures) than low-frequency features (e.g., colors), with theoretical and empirical justifications. To better restore hazed high- and low-frequency features, we decompose the hazed images into high- and low-frequency feature components with spectral dual-channel encoding and respectively design effective neural network architectures to recover hazed images on the two feature components. To be specific, we recover the low-frequency feature components with an encoder-decoder, while we specially design a high-frequency aggregation component (HFAC) to recover hazed images on high-frequency feature components, by referring to neighboring feature distributions. We conduct extensive experiments on four real-world image dehazing benchmarks. The experimental results show that our proposed SDCE framework outperforms the state-of-the-art baselines significantly, with an average 4.4% improvement in PSNR and an average 7.7% gain in SSIM.

Path inference aims to reveal missing paths given a few number of GPS samples associated with a moving object by exploiting the topology of road network and statistical information of historical GPS trajectories, and plays a vital role in data preprocessing of location based information services. But, in practice path inference severely suffers from the data sparsity as well as the randomness of drivers path selection behaviors. In this paper, we propose a novel Bayesian path inference model subject to spatiotemporal constraints by taking into account the drivers path selection behaviors. To be specific, the problem of path inference is cast as the problem of searching K most probable candidate paths according to the joint posterior selection probabilities of candidate paths. When estimating model parameters, we use the frequency of each road segment in the historical GPS trajectories instead of that of road segment transfers to mitigate the influence of data sparsity. In addition, both spatiotemporal constraints and probability thresholds are introduced to narrow the search space, which significantly improves the time efficiency. The experiments are conducted using practical data and show that the proposed model is significantly superior to three existing popular models. When the GPS sampling interval varies from 1 minute to 5 minutes, the accuracy of the proposed method is 0.94, 0.91, 0.86, 0.80 and 0.74, and the Jaccard similarity 0.89, 0.85, 0.83, 0.80 and 0.75 respectively, the average improvement in accuracy rises from 3.68% to 18.69% and that in the Jaccard similarity from 4.56% to 18.42%.

Identifying the root cause of congestion and taking appropriate strategies to improve traffic network performance are important goals of Advanced Traffic Management Systems (ATMS). On many occasions, the causes of congestion are not necessarily attributable to road infrastructures themselves. Instead, signal control strategies at intersections are very often the major contributors of congestion. In lieu of this, in this paper, a root cause identification method is developed with consideration of the impact from both road infrastructure and traffic signal control. Firstly, we differentiate congestion effects between road segments and intersections to attribute the causes of congestion to road infrastructure and signal control respectively. Then, we construct causal congestion trees to model congestion propagation and quantify congestion costs for each road segment and intersection in the whole road network. A Markov model is utilized to capture congestion spatio-temporal correlation among multiple road segments and intersections simultaneously, with which the most critical root cause can be located. Furthermore, a gradient boosting decision tree based method is presented to predict the root cause of congestion according to traffic flows, signal control strategies and road topology in traffic networks. Finally, simulations based on Simulation of Urban Mobility (SUMO) validate the effectiveness of our proposed method in identifying and predicting the congestion root cause. Experiments are further conducted using inductive loop detector data to identify the root cause for the road network of Taipei.

Motivated by the network tomography, in this paper, we present a novel methodology to estimate link travel time distributions (TTDs) using end-to-end (E2E) measurements detected by the limited traffic detectors at or near the road intersections. As it is not necessary to monitor the traffic in each link, the proposed estimator can be readily implemented in real life. The technical contributions of this paper are as follows: First, we employ the kernel density estimator (KDE) to model link travel times instead of parametric models, e.g., Gaussian distribution. It is able to capture the dynamic of link travel times that vary with the change of road conditions. The model parameters are estimated with the proposed C-shortest path algorithm, K-means-based algorithm, as well as expectation maximization (EM) algorithm. Second, to reduce the complexity of parameter estimation, we further propose a Q-opt and an X-means -based algorithm. Finally, we validate our proposed method using a dataset consisting of 3.0e +07 GPS trajectories collected by the taxicabs in Xi'an, China. With the metrics of Kullback Leibler and Kolmogorov-Smirnov test, the experimental results show that the link TTDs obtained from our proposed model are in excellent agreement with the empirical distributions, provided that ~70% of the intersections are equipped with traffic detectors.

This paper proposes a unified spatio-temporal model for short-term road traffic prediction. The contributions of this paper are as follows. First, we develop a physically intuitive approach to traffic prediction that captures the time-varying spatio-temporal correlation between traffic at different measurement points. The spatio-temporal correlation is affected by the road network topology, time-varying speed, and time-varying trip distribution. Distinctly different from previous black-box approaches to road traffic modeling and prediction, parameters of the proposed approach have physically intuitive meanings which make them readily amendable to suit changing road and traffic conditions. Second, unlike some existing techniques that capture the variation of spatio-temporal correlation by a complete re-design and calibration of the model, the proposed approach uses a unified model that incorporates the physical factors potentially affecting the variation of spatio-temporal correlation into a series of parameters. These parameters are relatively easy to control and adjust when road and traffic conditions change, thereby greatly reducing the computational complexity. Experiments using two sets of real traffic traces demonstrate that the proposed approach has superior accuracy compared with the widely used space-time autoregressive integrated moving average (STARIMA) and the back propagation neural network approaches, and is only marginally inferior to that obtained by constructing multiple STARIMA models for different times of the day, however, with a much reduced computational and implementation complexity.

User association has emerged as a distributed resource allocation problem in the heterogeneous networks (HetNets). Although an approximate solution is obtainable using the approaches like combinatorial optimization and game theory-based schemes, these techniques can be easily trapped in local optima. Furthermore, the lack of exploring the relation between the quality of the solution and the parameters in the HetNet [e.g., the number of users and base stations (BSs)], at what levels, impairs the practicability of deploying these approaches in a real world environment. To address these issues, this paper investigates how to model the problem as a distributed constraint optimization problem (DCOP) from the point of the view of the multiagent system. More specifically, we develop two models named each connection as variable (ECAV) and each BS and user as variable (EBUAV). Hereinafter, we propose a DCOP solver which not only sets up the model in a distributed way but also enables us to efficiently obtain the solution by means of a complete DCOP algorithm based on distributed message-passing. Naturally, both theoretical analysis and simulation show that different qualitative solutions can be obtained in terms of an introduced parameter which has a close relation with the parameters in the HetNet. It is also apparent that there is 6% improvement on the throughput by the DCOP solver comparing with other counterparts when . Particularly, it demonstrates up to 18% increase in the ability to make BSs service more users when the number of users is above 200 while the available resource blocks (RBs) are limited. In addition, it appears that the distribution of RBs allocated to users by BSs is better with the variation of the volume of RBs at the macro BS.