ABSTRACT Key Performance Indicator (KPI) anomaly detection is a crucial component of Artificial Intelligence for IT Operations (AIOps). In practical scenarios, service anomalies can lead to missing KPI values, which reduce the accuracy of anomaly detection. Moreover, existing methods frequently struggle to effectively detect subtle anomalies, thereby limiting detection performance. This study proposes a KPI anomaly detection framework that integrates missing value interpolation and adversarial training. First, a missing value interpolation model is built using an improved Transformer. The model is optimized via a mean absolute error loss constructed by randomly masking observations, while a diagonal mask in the self‐attention mechanism enhances interpolation accuracy. Second, a detection architecture is developed based on Variational Autoencoder‐Gated Recurrent Unit (VAE‐GRU) and adversarial training, which amplifies reconstruction error sensitivity to subtle anomalies. Finally, the Streaming Peaks Over Threshold (SPOT) algorithm is incorporated for adaptive thresholding based on the distribution of reconstruction errors. Experimental results demonstrate that the proposed framework achieves superior F1 scores and recall compared to existing methods on multiple benchmark datasets.

In the fifth generation mobile networks(5G), D2D (device-to-device) communication technology is introduced to provide services to users by utilizing the resources of idle mobile devices, and users can schedule computing tasks to be executed on mobile servers (MSs). This paper aims at reducing overall time delay and energy consumption by taking into account MSs mobility, task attributes, resource status for D2D scenario. We formulate the matching problem as a one-to-one matching game and propose a scheduling strategy based on binary graph matching algorithm. Simulations show that the strategy proposed in this paper can comprehensively reduce the task execution delay and energy consumption, and improve the success rate of task execution.

In Mobile Edge Computing (MEC), the services that a user receives change dynamically with location due to the user's mobility. If we mine the user's location data, predicting the user's next location, we can get the user's services to be used. It is convenient for the edge server to preload the user's services. When users reaches predicted location, the edge servers near users provide timely services. Therefore, this paper proposes a Spatio-temporal Position Prediction Model (SPPM) for Mobile Users Based on LSTM (Long Short-Term Memory) model in the mobile edge computing. Firstly, the time series feature extraction method is used to preprocess the historical location data of the mobile user. Next, the model uses the PCA data dimensionality reduction algorithm to process the data and then uses the LSTM model to predict the next spatiotemporal trajectory point of the mobile user. Finally, using the 17621 user trajectory data of the Geolife GPS trajectory data set, the algorithm is tested and verified. The experimental results show that the SPPM model proposed in this paper has higher prediction accuracy and more accurate prediction position.

In the edge computing, the service is deployed to the edge server through virtualization technology. Most strategies of service deployment are singleness and ignore the diversity of users' requirements and the service deployment cost at the edge. This paper proposes a clustering-based algorithm for service deployment, which considers the delay at the user side and the edge-side services deployment cost, and establishes a service deployment model based on multi-objective integer linear programming. Firstly, the K-means clustering algorithm is optimized to solve the problem of hotspot migration in the process of service deployment and reduce the deployment cost at the edge server side, then alternative enhanced heuristic algorithm is proposed to find the approximate optimal solution of services deployment. Experiments show that the algorithm can reasonably deploy services. Compared with traditional heuristic algorithms, the algorithm proposed in this paper has better performance in terms of user side and edge server side.

In order to reduce the delay and energy consumption of mobile devices, a computational offload strategy is adopted in mobile edge computing (MEC). At present, most computation offloading strategies only consider two computing resources, mobile devices and MEC servers. However, the computing power of the cloud server is much larger than that of the MEC server. Tasks with high computational complexity still need to be handed over to the cloud server for processing. This paper proposes an edge-cloud collaborative multi-task computing unloading model that considers both latency and energy cost. Usually the model solving is transformed into a search solution in finite strategy space. In this paper, the nonlinear exponential inertia weight particle swarm optimization (PSO) algorithm is used to get solution. By dynamically adjusting the inertia weight, the algorithm can make up for the convergence premature defect of the standard particle swarm optimization algorithm, and effectively avoid falling into the local optimal solution. Simulation experiments show that the strategy obtained by the model has lower total cost compared with different computation offloading models and strategies.

Semantic labeling in high-resolution aerial images is important for its wide range of applications. In this paper, we propose an end-to-end dual multi-scale feature fusion network (DMFFNet) for high-resolution aerial multi-source images. DMFFNet aims to further improve the semantic labeling results of the region where the multispectral features are indistinguishable. Specifically, we design a channel fusion strengthen (CFS) module, which can fuse features adaptively by modelling interdependencies between channels. Furthermore, a multiscale context aggregation (MCA) module is utilized to obtain larger receptive field and more contextual information. The experiment results confirm the DMFFNet with CFS and MCA improve the semantic labeling performance by utilizing multi-source data.

Semantic segmentation in high-resolution aerial images is a fundamental research problem in remote sensing field for its wide range of applications. However, it is difficult to distinguish regions with similar spectral features using only multispectral data. Recent research studies have indicated that the introduction of multisource information can effectively improve the robustness of segmentation method. In this letter, we use digital surface models (DSMs) information as a complementary feature to further improve the semantic segmentation results. To this end, we propose a lightweight and simple DSM fusion (DSMF) branch structure module. Compared with the existing feature extraction structures, proposed DSMF module is simple and can be easily applied to other networks. In addition, we investigate four fusion strategies based on DSMF module to explore the optimal feature fusion strategy and four end-to-end DSMFNets are designed according to the corresponding strategies. We evaluate our models on International Society for Photogrammetry and Remote Sensing Vaihingen data set and all DSMFNets achieve promising results. In particular, DSMFNet-1 achieves an overall accuracy of 91.5% on the test data set.