AUTOMATIC CONTROL AND COMPUTER SCIENCES最新文献

The research aims to propose a social network privacy protection scheme that combines differential privacy and multilayer grouping consensus algorithm to solve the problems of user privacy leakage and data abuse. Firstly, a community discovery data storage mechanism based on regional chains was designed to protect the security and integrity of data. Then, a multilayer grouping consensus algorithm was proposed to improve consensus efficiency through classification and hierarchical consensus. These results confirm that the proposed privacy protection scheme has improved privacy protection by about 75% and increased data availability by about 55% compared to other schemes such as Spctr Switch. When nodes are 200, compared to the traditional Byzantine consensus algorithm, the communication cost based on multilayer grouping consensus algorithm is saved by about 89.9%, and the consensus delay is reduced by about 75.6%. This research plan not only ensures user privacy and security, but also improves data availability, providing an effective method for social network privacy and security management, which helps maintain the stability and security of blockchain networks.

With the rapid development and progress of intelligent technology, interactive painting is facing unprecedented opportunities and challenges. The development of intelligent technology in the field of human–computer interaction has broken through the limits, making the interactive process more humanized, intelligent and diversified, but its creation mechanism has not yet been constructed and is in urgent need of theoretical guidance and normative constraints. To improve the adaptive image generation effect of art painting and enhance the efficiency of painting teaching, this paper proposes an interactive painting teaching system based on artificial intelligence technology. The innovation of this article lies in proposing an improved deep learning based image description generator model on the basis of existing image description generation algorithms based on adaptive attention mechanisms. The model is used to design a block sentinel that controls entity block switching and uses a two-layer LSTM with an improved adaptive attention mechanism as the image description generator. This method effectively improves the efficiency of image generation and enhances teaching efficiency. The experimental results show that the method in this paper can generate image contents accurately.

Aiming at large delay characteristics of the temperature control system, a radial basis function (RBF) method of superior and inferior mutation crossover strategies with storage mechanism differential evolution algorithm (SIMCDE) is proposed to tune and optimize the PID controller. The differential evolution algorithm introduces superior and inferior mutation strategies with storage mechanisms and superior and inferior crossover strategies, effectively avoiding the local optimal solutions. Then, the SIMCDE algorithm optimizes the initial parameters of RBF, and the gradient information is obtained by RBF online identification. Finally, three parameters of PID are adjusted online according to gradient information. Solving four test functions shows that the SIMCDE-RBF algorithm has good optimization ability. The simulation experiment of SIMCDE-RBF algorithm tuning PID parameters and the test of temperature control of heating furnace in a dairy company show that compared with IDE-RBF-PID, GODE-RBF-PID, and MCOBDE-RBF-PID, SIMCDE-RBF-PID has better dynamic performance, more robust anti-interference performance, and higher control accuracy.

As intelligent transformers continue to advance—a comprehensive preventive maintenance system has been gradually established for transformers. However, the main impediment to effective analysis is the imbalanced distribution of character data and positive anomaly data in the monitoring data of transformer oil, which adversely affects the intelligent evaluation of transformer status. Therefore, in this paper, we proposed analysis and research of unbalanced transformer insulation oil monitoring data using machine learning methods. First, we collected data pertaining to the status of insulation oil in smart transformers. Subsequently, it designed a numerical method based on word vector clustering tailored to the characteristics of insulation oil status data. Furthermore, a novel algorithm named KASMOTE (k nearest neighbor average smote) was introduced to process imbalanced insulation oil data. Finally, the paper validates the efficacy of the implemented dataset by employing seven machine learning algorithms. The experimental results demonstrate that the insulation monitoring dataset, incorporating word vector clustering and the KASMOTE algorithm, is both efficient and challenging, thus enhancing the feasibility of big data analysis.

The incorporation of renewable energy (RE) in the economic load dispatch problems (ELDPs) is not an easy task. This paper presents a reliable approach to solve the ELDPs of the hybrid renewable energy system (HRES) that consists of thermal, wind, and solar photovoltaic (PV) generators. The generation cost of RE is negligible, but the renewable operators demand some charge so-called renewable/maintenance/payback cost to run the plant. Therefore, the linear cost function has been implemented for RE generations (REGs). Two cases have been considered based on the cost of REGs, one is no cost for REGs and other is linear cost functions for REGs. The popular optimization technique known as particle swarm optimization (PSO) has been adopted to solve the ELDPs. A test system made of IEEE 30-bus system, solar PV, and wind generator has been considered to investigate the strength of the proposed approach. The simulation results show that the saving of 63.829, 74.99, and 182.937 $/h in one case and the saving of 139.53, 150.468, and 358.883 $/h in another case in the generation cost for a load demand of 283.4 MW are remarkable, when only solar PV, only wind and both solar PV and wind respectively, are in operation.

Most of the intelligent traffic light (ITL) devices are deployed in public places which are easily accessible by intruders that results various types of attacks. In order to secure the communication between ITL components, hierarchical blockchain based ITL (HBbITL) architecture has been proposed in this work. Proof-of-work (PoW) consensus has been implemented to enhance ITL device security. Elliptic curve digital signatures are employed for integrity of information and efficient processing in constrained resource environments. Among two popular Ethereum consensuses, PoW and proof of authority (PoA) implementation has been carried out for 1, 5 and 10 s, PoA provides higher throughput compared to PoW. Also, it is clear that HBbITL performs better than the public Ethereum blockchain with respect to latency.

Governments are actively seeking solutions to address the growing issue of longer waiting times for patients. To reduce the strain on the public sector and its increasing workload, the governmental bodies have established collaborative agreements with private healthcare service providers. While the private sector is expanding, it is not growing rapidly enough to meet the rising demands for healthcare services. Consequently, there is a dire need to explore innovative management techniques aimed at reducing patient wait times, cutting costs, and enhancing the quality of healthcare. In this paper, we propose an innovative solution to tackle the patient classification problem (PCP) using the machine learning paradigm. The proposed approach involves a hybridization of two classifiers, one utilizing the aggregation method and the other employing the support vector machine technique. We compare classification algorithms, including KNN, SVM, SVM + AM, and logistic regression, and evaluate their performance in terms of precision, recall, specificity, F1-score, and overall accuracy. The SVM + AM is found to be the best model for the classification of patients, followed by SVM, KNN, and logistic regression. We believe that such an evaluation will help addressing the challenges associated with patient classification, the medical practitioners, and, in turn, contribute to the overall healthcare system.

one of the important features of information retrieval systems is ranking. Ranking performs the function of ranking search results based on relevance to the user’s query. Methods developed in state-of-the-art research still require multiple iterations. In this paper, we proposed to use zone partitioning strategies for computing web page rank parameters in retrieval systems, which implements iterative calculation for only some randomly selected subgraphs (zone). The zone approach is based on the idea to use multiple neural networks to classify rank data in graph-based structures. The crawled web pages are fragmented into three distinct zones. The core zone is used for training graph convolutional network, in this zone, the labels are known. It is covered with an undiscovered zone, where classifiers label node parameters. The most interesting part is the intersection zone, which represents the set of nodes and edges that belong to more than one undiscovered zone. The experiments show that the probability of classifying the true labels in the intersection zones via aggregating the results of multiple classifiers in some cases is higher than in undiscovered zones.

In order to meet modern requirements for the development of socio-economic problems, it is necessary to develop and improve forecasting models. Existing fuzzy time series (FTS) forecasting models are built on the basis of the theory of fuzzy logic type 1, but the theory of fuzzy logic type 2 shows greater coverage of subject areas and more accurate modeling of the state of objects and systems. This is important because in reality the degree to which an element belongs to a particular set cannot be determined precisely, but only within a range. This paper proposes a fuzzy time series forecasting model based on the theory of fuzzy logic type 2 and the structure of Hedge algebra. The parameters of the proposed model are optimized using genetic algorithms. The proposed model is tested on the forecast of daily values of the Taiwan Stock Index (TAIEX) data, and the forecasting performance is assessed using the metrics RMSE, MAPE and MSE.

Unmanned aerial vehicles (UAVs) have made a significant impact on both industry and academics due to their many applications. It is convenient to exchange data, decision making, and control attitude/altitude of UAV systems with the establishment of a cloud communication network. There is a chance of data packet dropout and data packet delay during cloud communication. In this proposed work, the stochastic model of networked UAV is developed with network induced delay and packet loss using Bernoulli random variables. Also, discrete event triggered technique is implemented in sensor node and controller node that restricts unuseful information. Cloud network bandwidth and energy consumption are decreased as a result of limited data transmission. A predicted measurement is used to handle cloud network inefficiencies and during untriggered scenarios. For the developed stochastic UAV model, an estimator/filter is developed using orthogonal projection methods. An anomaly detection algorithm is proposed for a networked UAV system using estimator information to identify the fault and cyber-attack.