Keywords: logistics process, digital control, decision-making system, multivariate optimization, simulation
The article discusses the features of building a decision-making system with multivariate optimization of the structure of digital control of the logistics process. The key feature is the integration of the process of transferring the results of activities of the objects of the organizational system participating in the logistics process, and information exchange between the digital platforms of the control center and objects. The variety of structural relationships between the levels of digital control determines the need to choose the optimal option for two sets of indicators that characterize, on the one hand, the efficiency of movement of material flows, and on the other, the efficiency of information exchange between digital platforms. Based on these features, the components of the decision-making system have been substantiated, which allows, on the basis of the requirements determined by the management center, to choose the optimal structural solution. Since the selection process is carried out on a set of performance indicators, its formalized description is presented in the form of multicriteria optimization problems. In this case, for the algorithmization of decision-making, it is proposed to use a simulation model of a multiphase queuing system, adequate to digital control of the movement of material flows between objects of the organizational system.
Keywords: logistics process, digital control, decision-making system, multivariate optimization, simulation
The geological environment parameters are widely used in the hydrocarbon deposits exploration and study. They store information about the rocks physical characteristics, the location of the reservoirs, the boundaries of the layers separation, etc. The numerical values of the parameters under consideration are obtained by conducting geophysical well logging (GWL), or by recalculating existing logging curves based on known petrophysical laws. Because GWL can be carried out only if there is a drilled well, the environment parameters are often known only in a small near-wellbore space. This makes urgent the task of predicting the geological environment parameters within the entire field. The existing methods for solving this task can be conditionally divided into two groups: using only GWL data; complexly using 3D seismic and GWL data. The first group is effective when there is a dense grid of downhole measurements. The second group shows the best results in the conditions of a sparse irregular grid, however, it requires the seismic data availability within the entire studied field. This paper describes a new method for predicting geological environment parameters based on the GWL data and 3D seismic prospecting combined use. The paper also provides an algorithmic implementation of the new method; ways of optimizing algorithmic implementation are considered.
Keywords: volumetric petrophysical model, geostatistics, kriging, seismic exploration, geological well logging, seismic attribute, optimization
This article considers the issue of ensuring the effectiveness of the use of proactive control of the functional characteristics of the radar complex, taking into account the technical readiness of hardware and software and the level of training of operators and maintenance personnel. The importance of the contribution of the human factor in ensuring the combat readiness and effectiveness of the use of the radar complex for its intended purpose is emphasized. As a result of the analysis of the existing methodological apparatus for assessing the level of training of operators and service personnel for the implementation of modern methods of proactive management, the necessity and ways of improving it are determined on the basis of the competence approach adopted a priori in the Russian Federation. The concepts of the professional competence portrait of the operator and the reference professional competence portrait describing the necessary and sufficient set of competencies required by the operator for timely and effective management and application of the radar complex are introduced. An information model of a professional competence portrait is described, a metric for calculating a quantitative integral assessment of the level of its formation is proposed, which is applicable for software implementation as part of automated training tools. The scheme of application of automated training tools for solving operational tasks of proactive control of the radar complex is presented. The necessity of using Big Data of operational data of the repair and diagnostic complex for detecting signs, determining the causes and eliminating/preventing failures, as well as archival and up-to-date data from automated training facilities on the state of training of operators and maintenance personnel in the process of proactive control of the radar complex is noted. Based on the results obtained, it can be concluded that it is advisable to integrate training facilities and systems of technical and functional control of the radar complex in order to achieve a given level of probability of performing the tasks of the radar complex according to its intended purpose.
Keywords: proactive management, technical control, artificial intelligence, big Data, educational and training tools, competence approach
Keywords: , , ,
Differential diagnosis of focal pancreatic pathology is a complex clinical task. The method of endoscopic ultrasonography is the most informative in the detailed visualization of the pancreatic parenchyma and is used as a clarifying method in the diagnosis of focal pathology of the organ, which is provided by high-resolution images due to the minimal distance of the sensor from the object under study. One of the informative criteria for the differential diagnosis of the pathology under study is the characteristic of the contours of focal formations. The aim of the work was to improve the quality of differential diagnosis of focal pancreatic pathology according to endoscopic ultrasonography based on the analysis of the characteristics of the contours of focal formations with an assessment of their informative value. The results of endoscopic ultrasonography of 109 patients with solid tumors of the pancreas and 40 patients with chronic pancreatitis were analyzed. The endoscopic video system of the Olympus EVIS EXERA II company with the EU-ME1 ultrasound processor was used. Using the developed universal algorithms, the most homogeneous fragments of the region of interest are selected from the primary video materials, on the basis of which a scale of characteristics of the contours of the considered focal pathology is formed. Eight main types of contours corresponding to chronic focal pancreatitis and ductal adenocarcinoma with various degrees of differentiation were identified. The obtained data allowed us to construct a histogram with subsequent statistical analysis of the results, which showed the prospects of the proposed method of differential diagnosis of pancreatitis and oncological diseases and the synthesis decision rules.
Keywords: endoscopic ultrasonography, characteristics of the contours of focal formations of the pancreas, chronic pancreatitis, ductal adenocarcinoma, statistical analysis
This paper deals with the management of network structures to support decision-making in a data transmission network in terms of the efficiency indicator, which in general form is a generalized indicator of the stability of the network functioning [1, 2]. Management in communication networks is regulated by the Law "On Communication". It is a set of measures of an organizational and technical nature, the focus of which is to ensure the functioning of communication networks. This includes traffic throttling, including bandwidth management, channel resource allocation, and limiting the use of certain applications if necessary. Nowadays, at the present stage of development of technologies, the management of communication networks looks like a process of observation (measures for monitoring the state of nodes, lines), the choice of a control action and its application. The interaction of nodes is also monitored, and running applications are managed. The article discusses the decision-making mechanism in such a management of communication networks, which is characterized by multistage. Communication network management dictates the adoption of this or that decision in the event that the performance indicators change. Dynamic programming is a mathematical apparatus that approximates the solution of a certain class of problems, breaking them into parts, small problems and less complex ones. The paper presents cases of basic efficiency and efficiency of the network during restructuring to a new operating mode. When building a management system for special communication networks, several aspects can be distinguished. The dominant among them are the architecture and structure of the control system, control areas, control methods, as well as the protocols used. period, the process of restructuring.
Keywords: situational management mode, the basic performance, the decision time period, the process of restructuring
To restore the motor functions of the lower extremities of post-stroke patients, it is proposed to use a biotechnical system with a robotic device. The control is based on the analysis and classification of electromyosignals. The robotic device is controlled by a fuzzy control module, which allows maintaining three modes of rehabilitation and selecting and switching them depending on the functional state of the patient, thereby selecting the optimal rehabilitation program for the current functional state of the patient. The control model for includes three fuzzy control modules with the corresponding bases of fuzzy decision rules and allows you to adapt the rehabilitation procedure to the functional state of the patient. To assess the effectiveness of the proposed method of rehabilitation, the experimental group included 23 patients who underwent exacerbations from 25 days to 5 years, including patients with subacute (<180 days after exacerbation) and chronic (> 180 days after exacerbation) conditions. After a course of rehabilitation by means of a biotechnical system with a fuzzy control module, there is a significant increase in the maxima of the support reaction force Rz on the affected leg in the experimental group in relation to the control group. Accordingly, the amplitude of the front push in the experimental group increased by 62% (120%), the rear push by 58% (115%), while in the control group the amplitude increase was 40% (101%) and 41% (105 %). In this case, distinct maxima of the support reaction component Rz appear on the paretic leg.
Keywords: biotechnical system, fuzzy control module, post-stroke patients, robotic device, algorithm, base of fuzzy decision rules, edectromyosignal
The article describes an algorithm for regulating the quantity of slots in the implementation of the anti-collision protocol Q when it is used in information systems for the international goods transportation. The purpose of the study is to develop and substantiate an algorithm for regulating the number of slots when implementing the anti-collision protocol Q in information systems for international goods transportation. Results. An appropriate algorithm was proposed, which differs from existing solutions in fewer required variables and fewer operations required to calculate the number of slots. This algorithm should help to decrease the time needed to transfer information from RFID tags to information systems for international goods transportation. Upon condition an equal quantity of frames, the proposed algorithm reduces the total number of operations to determine the number of slots that will be required to read a set of RFID tags by 16.7%. The algorithm is applicable due to the fulfillment of the requirement to achieve maximum system efficiency not lower than for existing solutions, which was proved with simulation modeling. Practical significance: the results obtained can be used to ensure the availability of information when information from RFID tags enters the information systems for the international transportation of goods, as state information systems. The results can be used by customs authorities in the development of information support for labeling and traceability systems
Keywords: RFID-technology, RFID-mark, RFID-collision, mark collision, RFID-marking, customs bodies, goods marking, RFID, anticollision protocol
Abstract: The challenges of modern information security threats that have a destabilizing effect on special purpose communication systems (SSSN) require a response. At the same time, counteraction in the SSSN is considered as a complex, multilevel hierarchical process, and the information conflict is considered as a way of interaction between the components of the system, the result of which is not predetermined in advance. As a result, it is possible to implement various strategies for protecting information systems based on possible options for action. When analyzing the information conflict in the SSSN, the models of conflicts based on control theory, game theory, and the probabilistic-temporal approach are considered. On the basis of the principle of direct modeling of the conflict process, a model of conflict interaction "СССН-intruder" is proposed. The proposed model of the conflict "СССН-intruder" at the conceptual level is structurally presented in the form of a system with two control loops, having a common object of influence СССН and directly opposite goals of functioning, and the dynamics of their interaction is presented in the form of a scheme "impact-response" with the possibility of making a decision on the application of measures to counter conflict impact both after the fact of impact, and after establishing the fact of active scanning of vulnerabilities. The counteraction strategy in this case involves the use of the available resource, according to a certain algorithm, for the response of the system to the impact. It is shown that the choice of the countermeasures strategy, as well as the conflict interaction itself, depends on the operating conditions and parameters of the SCCH. As a result of modeling, it was shown that the conflict interaction over the entire time interval is cyclical, i.e. there is a predominance of the conflict impact or the strategy of counteraction to the conflict impact, at each step of the conflicting interaction of the parties.
Keywords: protection strategies, threats to information security, communication system of special purpose, information conflict, conflict interaction, management of the information security system
The article is devoted to the problem of detecting network attacks in Industrial Internet of Things systems. The topicality of the problem under consideration due to a high level of security risks in such systems is analyzed. Various algorithms of network attack detection are considered, and an increasing interest to applying methods of artificial intelligence for solving this kind of problems is noted. The advantages of combining various algorithms of artificial intelligence and methods of machine learning as a part of hybrid intrusion detection systems are underlined. The approach to design of hybrid intelligent intrusion detection system (IDS) is proposed, which includes at the lower level the artificial immune system, responsible for detection of anomalies and unknown network attacks, fulfilling so a function of preliminary network traffic filtration, and the multiclass classificator at the upper level, determining the class of the attack detected at the lower level of the system. The neural network and the random forest algorithm are considered as methods of constructing the classifier of the upper level. The training and efficiency estimation of the system proposed were carried out with use of the NSL-KDD dataset. As experiments showed, the best results were achieved by combination in hybrid IDS of the algorithms of artificial immune system and random forest.
Keywords: information security, network attack, machine learning, artificial immune system, neural network, random forest, hybrid intelligent system
A two-person noncooperative game that models a process of purchasing protection means for a computer system is considered. One of the players in this game is a party responsible for the security of the system. Having a certain amount of money that can be spent on the purchase of the protection means this party determines which of these funds should be purchased. Actions of the other player (and it's the external world in relation to the computer system) are attacks on the computer system implemented via the network. For each of the protection means that can be purchased as well as for each of the types of attacks that can be used in an assault on the computer system a probability with which the attack will be reflected by the protection mean is known. By choosing the protection means a party responsible for the security seeks to minimize overall losses which include first a cost of the purchased protection means and secondly a damage expected from use of the other party attacks on the computer system. A study of an optimality principle implementations of which are lexicographically maximin strategies of a player which is a party responsible for ensuring the security of the system is carried out. A result of this study is statements that determine a method of finding all lexicographically maximin strategies of the specified player.
Keywords: noncooperative game, maximin strategy, nondominated strategy, lexicographically maximin strategy, computer system, attack on a computer system, protection of a computer system
The scientific idea of this work is the formalization of the known thesis that there is a direct relationship between the quality of project management and consumer properties of the resulting product. Hardware and software complexes are a kind of complex systems, which proves the possibility of scientific adaptation of well-known approaches used in the research of complex systems of a different nature to the field of ensuring the functional safety of the hardware and software systems. One sign of this "relationship" of software and hardware systems with complex systems of different nature are typical problem situations arising in the implementation of projects. In the publications, at the declarative level, typical problem situations that occurs in managing of complex systems of different nature are described, and are called system archetypes. In this paper, on the example of the system archetype “limits of success’, a formal approach to the analysis of the compliance of the quality of hardware and software complex project management a with the state of the control object is proposed.The paper proposes a new systemic combination of models and approaches known in system analysis, which made it possible to increase the degree of formalization of research focused on assessing the impact of project management quality on the functional safety of hardware and software systems.
Keywords: project management, problem situation, functional safety, hardware-software system, system archetype “limits to success”
The research is devoted to the construction of a convolutional neural network model for recognizing medical images on the example of X-ray images database of patients with an established diagnosis of brain tumor. The convolutional neural network model is proposed, the architecture of which includes two convolutional layers and one fully connected layer. The accuracy results of the proposed classifier and accuracy results of the pre-trained models VGG16, VGG19, Inception-V3, InceptionResNet-V2, ResNet50, ResNet152 and Xception are compared. The considered CNN models on the test dataset achieved the image recognition accuracy from 95.36% to 98.84%. The highest accuracy of the results in solving the problem of recognizing a brain tumor was achieved by the models VGG 16, VGG 19, Xception and the proposed model. However, the training time of the constructed models differs depending on the architecture of the neural network. At the same time, for the proposed CNN model, there 0.783% were no detected signs of the disease among the X-ray samples of patients with an established diagnosis of a brain tumor. The proposed neural network model can act as an additional tool of a doctor in the diagnosis of a brain tumor. The introduction of computer vision algorithms into the daily work of a doctor will make it possible to promptly carry out an additional examination of the patient, make a diagnosis and carry out treatment in a timely manner. The use of services based on artificial intelligence algorithms can reduce the total time spent on diagnostic studies, identify pathologies at an early stage of the disease and are more likely to expect that treatment will lead to positive results.
Keywords: convolutional neural networks, image recognition, image classification, brain tumor, pretrained neural networks
Population algorithms enable simultaneously search many elements of approximation of Pareto optimal decisions set and hereupon provide large advantage in consumption of time compare to scalar goal function method that found a single decision in the search cycle. The capability of open-source platform PlatEMO for solving of problems of multiobjective optimization of electrical filters characteristics was investigated in this work. Experience has shown that for two-objectives optimization problems only 6 algorithms of 71 provided good results. Approximations of Pareto set found by these algorithms were better than approximation found by scalar goal function method. Comparison was carried out by means of Coverage indicator that estimates the part of the first approximation elements dominated by the second approximation elements. For three-objectives optimization problems only two algorithms provided acceptable results. In this case approximations of Pareto set found by population algorithms were worse than that found by scalar goal function method. The conclusion was made that a rational method may consist of application of population algorithm for the solving of several two-objective optimization problems with constrains on other objectives and successive aggregation of found subsets.
Keywords: pareto-optimality, population algorithm, scalarization, decomposition, dominance, gain-frequency response, phase-frequency response
Abstract: One of the most important problems in the organization of an effective educational process is the problem of matching the qualities of teachers and students. The relevance of solving this problem is determined by the growth of the volume and complexity of educational material, as well as the introduction of digital technologies into the educational process. Digital technologies allow us to formally describe the characteristics of the personalities of teachers and students who have the greatest influence on the educational process. These properties of personalities form their portraits. The purpose of the article is a vector-stochastic description of a joint portrait of a teacher and a student. To achieve this goal, the apparatus of formal grammars is used. It is proposed to associate a vector of a point assessment with each student and each teacher. The vector evaluation is presented in the form of deducible chains of a formal language in a probabilistic grammar. The corresponding output tree is built. The output tree is considered as a network of a special type – an information graph. Matrices describing the graph are constructed (the adjacency matrix and the matrix of the sum of adjacency matrices in different degrees). The resulting graph describes the information system. It is shown how to calculate the probabilities of finding the system in each level. To explain the developed method, a specific numerical example of calculating the system parameters is considered. The criterion of the dynamic unity of the portraits of the teacher and the student is formulated. The result of the research is the developed new method of mathematical description of the joint portrait of a teacher and a student. The developed method can be used in models of digitalization of education and similar areas.
Keywords: evaluation vector, inferred language chains, probabilistic grammar, rule scheme, inference tree, information graph, system, adjacency matrix
One of the most important factors that are affecting of quality of processes service in a distributed system is the way of managing shared resources. At the same time, it is necessary to take into account that each process (task) has a set of characteristics, the complex accounting of which makes it possible to increase the efficiency of executing processes. Among the most important characteristics are the execution time of the process and its significance for solving system-wide tasks. This article is devoted to the development of a resource allocation algorithm based on a two-criteria process assessment. The priority and order of tasks execution is determined based on the importance weights formed by the PROMETHEE II multicriteria decision-making method. The paper describes the features of the application of this method to solve the problem and design an algorithm for the resource’s allocation based on a two-criteria assessment of processes. The algorithm provides for the possibility of interrupting the service of processes and forming a queue based on the importance weights. To automate the resource planning process, a software product has been developed that implements the stages of the algorithm. The calculations have shown that the proposed algorithm improves the quality of management of distributed systems, making the resource planning process more flexible and efficient. The approach described in the work is universal and can be extended for the case of an arbitrary number of criteria for evaluating processes.
Keywords: distributed systems, resource allocation, multi-criteria decision-making (MCDM), PROMETHEE II, virtual machine
Currently, simulators are used in various fields of training specialists. Especially the ability to repeatedly and accurately recreate various situations allows students to practice a variety of scenarios. In medical education, there is a need to improve the acquired knowledge in order to prepare for working directly with patients. It is not always possible to conduct practical classes, and even the proposed practice cannot cover the entire range of clinical cases. One of the most important aspects of training a doctor is the development of certain thinking skills, and the closer they are to real practice, the more effective the training of a specialist will be. The automated training system is a complex of technical, educational, methodological, linguistic, software and organizational support on a computer basis and provides a simplified model of the phenomenon being studied. This facilitates different cases easy perception and research, the development of certain skills and abilities to apply special knowledge. The article identifies the some problems of diagnosing abdominal diseases. The solution to these problems can be the introduction of an auxiliary tool in the form of an automated training system at the stage of training specialists. A database for the system operation is proposed, the order of its formation and building is described. Flowcharts for drawing up a task variant are developed. Also an interaction workflow and a possible interface forms design of the system modules are presented.
Keywords: diagnostic errors, medical training simulator, abdominal pathologies, situational tasks, clinical thinking, medical education, automated training system
For many years, assisted reproductive technologies (ART) have been helping to conceive a child when this is not possible naturally. We can consider the ART protocol to be successful not only in case of pregnancy, but also in case of its successful completion: the birth of a healthy child. The article describes the creation of a software application for employees of ART centers, which helps to predict the outcome of the protocol, including the probability of pregnancy, the forecast of possible complications during its course, the forecast of the time and method of delivery, and the health group (1-5) of the born child. To create the application, we used data on 854 protocols implemented in 2016-2018, because of which 464 children were born. The analysis of their health contains information from birth to three years of age. The application uses sixteen binary classifiers, nine of which implement multi-class classifications of the term of delivery, the delivery method and children’s health groups. To implement multiclass inference, the “one-vs-all” strategy was used. Сross-validation was used to check the quality. The remaining 7 classifiers predict the likelihood of pregnancy and the occurrence of its complications: cervical incompetence, hypertensive disorders, placenta previa, gestational diabetes mellitus, violations of the amount of amniotic fluid and premature rupture of the membranes. We have built all the models based on the random forest algorithm using the Python programming language. The interface was created using the PyQT5 and QtDesigner libraries.
Keywords: machine learning, assisted reproductive technologies, expert system, software application, child health status prediction
The issue of constructing a solution to an initial-boundary value problem for an evolutionary differential equation with a spatial variable varying on a network (geometric graph) has remained under review of researchers over the past few years. There were many practical reasons for this - a large number of mathematical models describing the transport processes of continuous media over network carriers use formalisms of partial differential equations and their corresponding initial-boundary value problems. In this paper, classical approaches were used for approximating differential equations on linear network fragments (graph edges), and the principles of constructing approximations of differential relations generated by generalized Kirchhoff conditions at the junction points of these fragments (at the graph nodes) were also indicated. The latter was a distinctive feature of differential equations concept with a spatial variable, changing on a network (graph) and its corresponding finite-difference analogue of the classical equation and finite-difference analogue. The problems of the elliptic operator approximation of the initial-boundary value problem were studied (the error of approximations was established), the stability of the two-layer difference scheme and detailed analysis of its stability was carried out. An algorithm for constructing a solution was developed, based on new numerical methods for analyzing transport problems in materials with complex structure with non-uniformly distributed properties of a continuous medium over a network carrier. A computer program has been developed and tested on test objectives targeted at applied problems. The obtained results can be used in the analysis of initial-boundary value problems for differential equations with distributed parameters on a multidimensional network having interesting analogies with multiphase problems in multidimensional hydrodynamics.
Keywords: initial-boundary value transfer problem, network (directed graph), weak solution, finite-dimensional analogue of differential operators, stability of difference schemes
The article deals with the problem of optimizing the educational process through the use of cloud technologies. The purpose is to study the effectiveness of using cloud technologies in solving problems of optimizing the educational process. The paper considers the theoretical aspects of optimizing the educational process, the features of cloud technologies; the analysis of existing cloud computing solutions in the educational space is carried out; the necessity of using cloud technologies in teaching based on computer classes is formulated; the results of using the cloud service for solving optimization problems in the educational space of the school are presented. An analysis of existing cloud computing solutions has shown that there are currently three models for deploying cloud systems: private, public and hybrid, and three service models depending on the type of services provided: IaaS, PaaS and SaaS. In accordance with the objectives of the study, Amazon Web Services was selected to achieve the research goal, at the level of free access for one year monthly for 750 hours of operation of the t2.micro instance with Linux or Windows. The results of the study showed that when using cloud technologies in combination with pedagogical technologies, it was possible to increase the level of knowledge and skills of students in Computer Science and information technologies with the least time, while minimal material resources were spent on achieving the set results for a certain period of time.
Keywords: educational process, training optimization, information technologies, cloud computing, service models, deployment models
The current-voltage characteristic (CVC) is an important integral characteristic of the salt ion transfer process in electromembrane systems, which are considered as the desalination channel of the electrodialysis apparatus. The article examines the theoretical current-voltage characteristic, for the calculation of which a new 2D mathematical model of non-stationary 1:1 transfer of an electrolyte in a potentiodynamic mode is formulated and numerically solved, taking into account the electroconvection and non-catalytic reaction of dissociation and recombination of water molecules. The main regularities of changes in the current-voltage characteristic and their connection with the electroconvection and non-catalytic reaction of dissociation and recombination of water molecules are established. It is shown that before the occurrence of electroconvection, the values of CVC, taking into account the dissociation/recombination reaction of water molecules, are higher than the values of CVC without taking into account this reaction. This difference is caused by the effect on the electric field strength of the products of water dissociation, i.e., the exaltation of the limiting current. Electroconvection begins later, taking into account the dissociation/recombination reaction of water molecules, than without taking into account this reaction. At higher values of the potential jump, the values of the VAC taking into account the dissociation/recombination reaction of water molecules are lower than the values of the CVC without taking into account this reaction. It is established that the non-solenoidal part of the current is small, so the total current and the solenoid part of the current coincide with good accuracy, both in the case of taking into account and in the case without taking into account the dissociation/recombination reaction of water molecules. Thus, in the first approximation, the total current can be considered as the solenoid part of the current, which is calculated using a double integral that is resistant to rounding errors in spatial variables, but retains all the features of the change in current density over time.
Keywords: current-voltage characteristic, membrane systems, mathematical model, cross-section of the desalting channel, ion-exchange membrane
In this article a neural network trained by hybrid neuroevolution solves the maze problem. Hybrid neuroevolution combines differential evolution with the novelty search. Algorithms that preserve the best solutions face the problem that estimates of the novelty of these archival solutions will not change from generation to generation. This article aims to address this problem by proposing two methods for adjusting estimates of the novelty of solutions: novelty destruction and actualization of novelty rates. The novelty destruction allows novelty to diminish over time, thereby allowing the search algorithm to evolve, while the actualization of novelty rates updates the novelty of these solutions in each generation. When testing on the problem of navigation in the maze, it was noticed that the novelty destruction and the actualization of novelty rates converge faster than just the standard search by objective function and the novelty search.
Keywords: neuroevolution, neural networks, maze, novelty search, differential evolution
The result of the solving of crimes is one of the important indicators of the activities of law enforcement agencies. Despite the improvement of crime investigation methods, the success rate of crime detection in the Russian Federation remains at the level of 51%–56%. The article describes a method for constructing a mathematical model – a digital double of a registered crime. As the initial data for constructing the model, an array of information – primary accounting documents, about 341 thousand crimes committed on the territory of the Primorsky Krai over 11 years-from 2010 to 2020. The model allows you: with 88% confidence, based on the formalized primary information contained in the primary accounting documents – statistical cards Form No. 1 “On the detected crime”, to make a forecast about whether the crime will be solved or not; to audit unsolved crimes of previous years in order to determine the crimes that have a high probability of detection; to identify the features in the statistical cards that most affect the forecast of the detection of crimes. The model is based on the use of machine learning algorithms “gradient boosting over decision trees”, implemented in the open library of artificial intelligence CatBoost from Yandex. The accuracy of the model is confirmed by the preparation and verification of the forecast of the result of the investigation of crimes in January–June 2021 for 16408 crimes committed on the territory of the Primorsky Krai.
Keywords: digital double, predictive model, crime, statistical cards, machine learning, artificial intelligence, catBoost, gradient boosting, decision trees, feature importance
The main trend in the development of modern technical means for the protection and supervision of objects of special importance (ALE) is the integration of various technical means and systems into a single information and technical complex - an integrated security system (ISS). ISF provide comprehensive security of objects of protection by integrating various subsystems into a single hardware and software (information) platform. The role of elements of these information systems may well be, for example, trainable complexes of automation of checkpoints (checkpoints), designed to ensure control and management of personnel and transport access to the restricted areas of the ALE in an automated mode. Currently, there is a problem of finding and choosing inexpensive automated control systems in technical systems of objects of particular importance with the implementation of integration algorithms with database management systems (DBMS). Modern access control systems are not always suitable for solving some of the tasks inherent in special-purpose facilities. This is due to the sometimes high cost of software or technical solutions, the limited functionality of devices from specific manufacturers, and the atypical nature of the tasks being solved. In other words, the development of problem-oriented control systems, information processing and decision-making in order to optimize the technical support of ALE is currently in demand.
Keywords: algorithm, control, information system, information process, programmable microcontroller, database management system
Currently, cloud computing technology is gaining popularity – a technology in which the computer resources of a data center are provided to the user over a network, as an online service. Demand drives the growth in the number and size of data centers that provide these services. The pandemic has led to the transition of many services to online mode, many organizations have emphasized the effectiveness of remote work, and the prevalence of distance learning is growing. Thus, there is a need to optimize the operation of the IT infrastructure of cloud service providers in order to increase its cost-effectiveness and environmental friendliness (compliance with the Green computing concept) while maintaining a predetermined level of service quality. One of the key challenges of providing cloud services is the optimal distribution of virtual machines on physical servers. This problem has been studied by many researchers, and in this work the analysis of existing approaches to its solution is carried out. All of them are based on analyzing the current workload and then adjusting the distribution of virtual machines. We propose the intellectualization of the use of hardware resources of the data center, which consists in the proactive management of hardware platforms, the placement of virtual machines, based on the prediction of the workload in the future. To test the proposed approach, we used the capabilities of the CloudSim framework.
Keywords: cloud computing, virtualization, workload forecasting, cloudsim, datacenter, optimization
The article examines one of the processes associated with management in a network organizational system based on a leadership strategy, the process of classification transformation of grouping objects according to the conditions of the management center. The main task is to determine the boundary values of the parameters of the conditions specified on a discrete set in such a way as to distribute resource provision to objects with the maximum potential for achieving target indicators. It is shown that such a problem, when formalized, is a Boolean programming problem. Optimization models of classification ordering are presented for two conditions for rating objects to top classes. These models support a first-order classification ordering process. For solving the task of the secondary sequences of objects not included in the top class, a characteristic feature of the stepwise variation of the integral function is used depending on the numbering of objects corresponding to the rank sequence. Separate stepwise changes are identified as function jumps and determine the optimization nature of their belonging to a given class. An optimization problem is a minimal coverage Boolean programming problem. Using a unified scheme based on the patterns of randomized search, genetic algorithm, and expert estimation is proposed to solve the generated Boolean programming problems.
Keywords: network organizational system, classification transformation, optimization modeling, boolean programming, integral estimates
А wide range of services based on artificial intelligence is expanding into a new industrial revolution. The article describes artificial intelligence technology, which is applied in real life due to the development of technologies related to deep learning and the ability to process large amounts of data. Chatbot with interactive artificial intelligence helps you shop online, get the latest news and information. The chatbot service, which some educational institutions have begun to implement, is only the first step towards automating the provision of services in the field of educational services. The article describes the administration of an interactive chatbot of the North Caucasian Federal University with artificial intelligence, presents the process of providing such services as remote submission of documents with downloading files of various formats, providing various information for applicants, advising students on accommodation in a hostel, etc. ... The study proves that chatbot services should be implemented in universities, and this proves the economic feasibility of development. The period of operation of the interactive chatbot of the North Caucasus Federal University is relatively short, therefore, at present, it is in the stage of continuous improvement.
Keywords: chat-bot, artificial Intelligence, intelligence Assistant, mobile Messenger, personal Assistant, system Integration
Artificial intelligence systems are used in many areas of human life support for example finance or medicine. Every year intelligent systems process more and more data and make more and more decisions. All these decisions have an increasing impact on the fate of people. The corner-stone is a distrust of completely non-human, autonomous artificial intelligence systems. The key to distrust lies in the misunderstanding of why intelligent systems make this or that decision, based on what beliefs such systems operate (and whether they have their own beliefs or only those that were given to them by the developers). To solve the problem of “distrust” in such sys-tems, the methods of explainable artificial intelligence have been used. This article provides a brief overview of the most popular methods in the academic environment such methods as PDP, SHAP, LIME, DeepLIFT, permutation importance, ICE plots. Practical exercises demonstrate the ease of application of PDP and SHAP methods, as well as the convenience of "reading" the graphical results of these methods, which explain the constructed decision tree model and ran-dom forest model on the example of a small set of sales data
Keywords: artificial intelligence, explainable artificial intelligence, interpretable artificial intelligence, explainability, interpretability, XAI, PDP, SHAP
It is shown that structural interference is one of the most dangerous for communication systems, for example, for radio security systems. Some of the well-known approaches for assessing noise immunity under the influence of structural interference are described. It is indicated that the analysis of the noise immunity of wireless data transmission systems under the influence of structural interference is analytically complex. Estimates of the noise immunity of coherent reception of signals with frequency and quadrature phase manipulations under the influence of structural interference are developed, based on the approach proposed in [6, 13]. Using the developed noise immunity estimates, graphs of the averaged dependence of the bit error probability on the signal-to-noise ratio are constructed. It is shown that coherent reception of signals with quadrature phase manipulation under the influence of structural interference at fixed values of the bit error probability provides more efficient signal reception compared to coherent reception of signals with frequency manipulation due to a lower signal-to-noise ratio. The obtained probabilistic values in the future will allow planning of measures, both for the establishment of structural interference, and for protection against them in radio security systems. In addition, the conducted research can be useful for developers and manufacturers of radio security systems. Further research in this area is promising, in particular, the use of complex noise-like signals, such as chaotic signals, to counteract structural interference
Keywords: radio security systems, manipulations, structural interference, noise immunity, assessing
The article presents approaches for identification of the author of a natural language text, the advantages and disadvantages of these approaches, describes the process of identifying the author of a Russian-language text using deep neural network architectures (LSTM, CNN with attention and Transformer). The relevance of the problem is due to the active digitalization of society and the transfer of most of the activities online. Authorship techniques are especially useful for information security and forensics. Such methods can be used for the identification of the author of suicide notes, as well as plagiarism detection. Plagiarism detection is an important problem for the protection of intellectual property in the digital space and also for the educational process. The results show that deep neural networks achieve an accuracy of 93% when determining the author of a text. The study also includes an evaluation of the impact of attacks on methods on the accuracy of the models. Experiments show that the loss of accuracy of deep neural networks does not exceed 20% when attacking the method, but SVM-based method is not resistant to deliberate anonymization of the text. The Transformer architecture is the most effective for anonymized text and achieves 81% accuracy.
Keywords: text mining, machine learning, attribution, neural networks, deep learning