With the fast growth of software-defined networking (SDN) and Internet of Things (IoT) ecosystems, the risk of getting major cyber-attacks has become a serious concern. As such, robust real-time intrusion prevention is drastically needed. In this paper, we present HashBoost-IPS, a novel intrusion prevention system that merges XGBoost-based attack detection with multi-hash cryptographic verification. The system detects malicious traffic in real-time, blocks it immediately, and ensures secure handling through cryptographic hash-based validation. The NSL-KDD dataset was used to test the system. The model works extremely well with efficient handling of features and almost no extra computing power needed. Findings show that HashBoost-IPS is an efficient and versatile solution for real-time intrusion prevention in SDN, IoT networks. Maintaining a low computational overhead, the proposed model was able to achieve accuracy of around 99.8 % . Besides, the model demonstrated high precision, recall, and F1-score.
The article describes the development and implementation of a microservice architecture upon which the functionality of an Intelligent Digital Twin of Plants (IDTP) is built. This system is designed for planning and simulation plant growth stages and forecasting yields, synchronized with the states of actual crops, based on ontologies and multiagent technologies. The functional capabilities of the Manager, Ontology Constructor and Knowledge Base (OC and KB), planning (multiagent planner) services are examined, as well as the organization of interfaces for the interaction of the IDTP with external services and users. Examples are provided to represent the results of modeling the comprehensive state of a crop, based on real-time environmental data and forecasting the crop’s condition at all stages of plant development up to harvest. The developed microservice architecture is open for integrating the IDTP with external services, such as environmental data collection, as well as recommendation systems and Agro-IoT platforms. The results of the IDTP prototype development have been passed on to several farms for experimental testing in different climatic conditions. The article discusses directions for improving the accuracy of IDTP forecasts and the potential for creating a decision support system for agronomic applications in precision agriculture based on the IDTP.
The development of a neural network for predicting the aerodynamic coefficients of airfoil sections for subsequent application in the blade element method to calculate the characteristics of aircraft engine propeller blades is described. A specific feature of this neural network is the ability to predict the dependences of the lift and drag coefficients on the angle of attack and Reynolds number of flows in the form of two-dimensional images of pixel color distributions. An airfoil parametrization method and the development of a database are presented, and the architecture of the proposed neural network is described. The results of training the neural network are reported, and its capability of predicting the aerodynamic characteristics is demonstrated.
The paper presents the results of developing a photocatalytic generator of chemically active oxygen species intended for air disinfection and purification in fruit and vegetable storage systems. A Cu/ZnO-ZnAl _2 O _4 composite oxide photocatalyst synthesized by the polymer-salt method was used as the active material; under excitation by 365 nm radiation, it provides efficient generation of singlet oxygen. The design of the generator is described; it includes preliminary mechanical air cleaning, a photocatalytic unit, carbon and paper filters, and a forced-air supply system. It is shown that the use of the device reduces weight loss in raspberries and improves the preservation of mango and avocado fruits during storage, which confirms the prospects of photocatalytic air purification for maintaining fruit freshness.
We consider the issues of creating algorithmic support for supervised classification problem, which is one of the central tasks of machine learning. Original procedures of logical analysis and classification of integer data represented as a set of elements of Cartesian product of finite partially ordered sets (product of partial orders) are constructed and investigated. At the training stage of the proposed procedures, the search for so-called regular representative elementary classifiers (special fragments in feature descriptions of precedents that distinguish objects belonging to different classes) is performed. An asymptotically optimal algorithm for enumerating the required elementary classifiers over a product of antichains is constructed and the results of its testing on real-world tasks are presented. Theoretical and experimental justifications for the efficiency of the new classification procedures are provided for the case when linear orders on sets of feature values are defined. The theoretical conclusions are based on the study of the metric (quantitative) properties of the set of regular representative elementary classifiers.
