Vol 14, No 4 (2023)
The organization of interaction the active objects of homogeneous digital structures
Abstract
This article extends the possibilities of interaction of active devices (objects) in homogeneous digital systems. The system consists of objects that are located within a few tens of meters and organize their interaction only by their own means. Objects can be stationary and mobile with arbitrary and time-varying mutual arrangement The homogeneity of the system means the absence of external control and equal opportunities of objects in the organization of their interaction. Communications between objects are wireless using optical or radio signals. Signals from any source object go directly to all receiver objects. The object receives the right to transmit signals deterministically according to its priority value, either dynamically or statically. The proposed structure and interaction methods allow objects to perform distributed group (associative) operations in addition to the usual distributed system message exchange. In them, objects simultaneously eliminate a group of conflicts of access to the common channel of data exchange, determine the state of all objects of the system, and synchronously perform joint actions of objects, reacting to unforeseen changing external conditions. A group of objects simultaneously participating in a group operation is selected by specifying a set of criteria that the objects must possess. The homogeneity of the system greatly simplifies its technical implementation, but the inclusion of heterogeneity in the system will accelerate the performance of group operations when becoming more complex. Therefore, the homogeneous system is proposed to be used mainly for communication between active peripheral devices and those with more complex cluster computers.
Program Systems: Theory and Applications. 2023;14(4):3-23
3-23
Method for classifying aspects of argumentation in Russian-language texts
Abstract
Argumentation mining in texts has attracted the attention of researchers in recent years due to a wide range of applications, in particular, in the analysis of scientific and legal texts, news articles, political debates, student essays and social media. Recently, a new task has been set in this area— aspect-based argumentation mining, where an aspect is defined as a property of the object, regarding which the argument is being built. Accounting for the aspects allows, on the one hand, to clarify the direction of the argumentation and understanding of the argument structure; on the other hand, it can be used to generate high-quality and aspect-specific arguments. The article proposes a method for classifying aspects of argumentation in texts in Russian. On its basis we train and study the models for classifying aspects of argumentation using machine learning and neural networks. For the first time, a Russian-language text corpus was formed, including 1,426 sentences and marked by 16 aspects of argumentation, a neural network language model ArgBERT for classifying arguments was built, and Random Forest models were trained to classify aspects of argumentation. The classification performance obtained on the basis of Random Forest models is 0.6373 by F1-score. The developed models demonstrate the best performance for the aspects “Safety”, “Impact on health”, “Influence on the psyche”, “Attitude of the authorities” and “Standard of living” (F1‑score is higher than 0.75).
Program Systems: Theory and Applications. 2023;14(4):25-45
25-45
Minimax improvement method for inhomogeneous discrete systems
Abstract
A class of two-level discrete inhomogeneous systems (DNS) is considered for the case when all homogeneous subsystems of the lower level are not only connected by a common functionality, but also have their own goals. Similar systems are widely used in practice (economics, ecology), and also arise in the process of numerically solving optimization problems when discretizing continuous control systems. A second-order control improvement method is proposed, for the derivation of which a generalization of sufficient optimality conditions by V. F. Krotov. Illustrative examples are given.
Program Systems: Theory and Applications. 2023;14(4):47-66
47-66
Cyclic pipeline systems
Abstract
One of the most efficient ways to organize calculations on ASIC or FPGA is the creation of non-stallable pipelines. However, for some computing circuits, the resulting pipeline may be too large for available ASIC or FPGA resources. The authors propose a method for constructing cyclic pipelines, in which data flow control is based on counters and does not depend on the data being transmitting along the pipeline. We proposed the method makes it possible to build more compact non-stallable pipelines. One of the main details of method is to use cycle ratio equal to the number of times the data must go through the loop, after which the pipeline converts the data into the desired result.
Program Systems: Theory and Applications. 2023;14(4):67-89
67-89
Coloured petri nets and the language for distributed programming UPL: their comparison and translation
Abstract
Petri nets are widely used as a means of modeling distributed multi-agent systems. There are tools for working with extended Petri nets, in which tokens are loaded with arbitrary data. For example, CPN Tools allows you to describe, play and explore Colored Petri Nets (CPN). The question is raised about the possibility of using this tool for the development, prototyping and research of parallel distributed computing algorithms, ideally turning them into working efficient parallel programs. We have experience in experimental programming of various algorithms in the graphical data flow language UPL, which currently exists “on paper”. Its comparison with CPN shows that their semantics have a lot in common. In the article, both languages are defined, compared with examples and through the rules of translation from one to another. The means for defining distribution of computations (distribution functions) are also described. An interesting question is about their transfer to CPN, where they have no analogues yet.
Program Systems: Theory and Applications. 2023;14(4):91-122
91-122
Generation of a plot component for a tabletop role-playing game
Abstract
The analysis of the subject area and relevance of tabletop role-playing games, as well as their impact on the spheres of human activity, is carried out. The mechanics of the gameplay of the tabletop role-playing game Dungeons and Dragons have been studied. Software solutions for creating a plot description of locations, including a graphical representation of maps and one-page adventures, are considered. Implemented design and adaptive user interface. The database architecture has been formed, as well as implemented in JSON file format. The architecture of the model has been developed, as well as its configuration. The block method of plot construction is described. The algorithm of the application for generating the plot component has been developed. Lists of plot components are formed, as well as the logic of random selection of sections of the description with a check on the adequacy and compatibility of the description is implemented.
Program Systems: Theory and Applications. 2023;14(4):123-140
123-140
Cloud service for differential diagnosis and personalized treatment of inflammatory heart diseases
Abstract
The aim of this research is to automate the process of differential diagnosis and prescription of personalised treatment of inflammatory heart diseases. The main method of research is the formation of declarative knowledge bases and a solver explaining its results. The ontology solver interprets the formalized knowledge. Main principles of development and the conceptual architecture of an intelligent service with declarative knowledge are described. Information components for inflammatory heart diseases and software components for the differential diagnostics and personalised treatment are formed. The clinical decision support system for cardiology with the properties of explainable artificial intelligence has been developed. It is hosted on the medical portal of the cloud platform IACPaaS. The practical significance of the results lies in the fact that the system is hosted on a platform with free registration. For the confidence of practitioners, knowledge sources are indicated and testing is performed on data described in the literature. The service is available for experimentation on the role of declarative knowledge base in real-life situations where a third opinion is required. The method used is also practically significant - the rationale for the choice of technologies and algorithms is given, the requirements for the software complex service are identified and formulated, and the steps for the development of all components are outlined. This allows us to reproduce the proposed solution on the same principles of explainable artificial intelligence, but with other knowledge and on another platform.
Program Systems: Theory and Applications. 2023;14(4):141-188
141-188
SDDRE based nonlinear feedback construction in the tracking problem for a wheeled robot model
Abstract
The article discusses the problem of constructing nonlinear feedback in the tracking problem for a wheeled robotic system. A special feature of the work is the formulation of a problem in which the reference trajectories of the system are known in advance, as well as a modification of a previously known algorithm based on the State-Dependent Differential Riccati Equation technique. Numerical experiments show that the proposed approach allows for a compromise between control quality and operating speed.
Program Systems: Theory and Applications. 2023;14(4):189-206
189-206