В работе представлены архитектура и инструментальное обеспечение цифровой платформы поддержки междисциплинарных научных исследований, разработанной в ИДСТУ СО РАН. Обоснована необходимость создания единой интегрированной среды, обеспечивающей сбор, хранение, обработку и публикацию научных данных на основе принципов открытой науки и FAIR (findable, accessible, interoperable, reusable). Предложена четырехуровневая сервис-ориентированная архитектура, включающая уровень инфраструктуры, уровень интеграции и управления данными, уровень базовых и инструментальных сервисов, а также прикладной уровень. Описан разработанный инструментарий: каталог данных и структурных спецификаций, конструкторы сервисов ввода и редактирования данных, отображения и визуализации пространственных данных, сервис импорта реляционных данных, а также компоненты сбора и анализа файловых данных. Показано, что предложенное решение позволяет унифицировать процессы создания сервисов, обеспечивает интероперабельность компонентов, масштабируемость инфраструктуры и снижает барьеры для междисциплинарного взаимодействия исследователей. The aim of this work is to develop an architecture for a digital platform to support interdisciplinary scientific research, which will overcome barriers related to incompatible data formats, fragmentation of information resources, and adherence to the principles of open science (FAIR: findable, accessible, interoperable, reusable). The proposed concept of the digital platform is based on a four layer service oriented architecture (SOA), comprising: the infrastructure layer (cloud resources, Kubernetes/JupyterHub containerization), the data integration and management layer (data lakes, metadata catalogs, ETL pipelines), the core and instrumental services layer, and the application layer (the domain specific user layer). Implementation leverages OGC standards (WPS, WMS), the OAuth 2.0 protocol, as well as visualization tools and service builders for working with both relational and spatial data. A digital platform has been developed and deployed, including a catalog of data and structural specifications, service builders for input, editing, display, and visualization of spatial data, along with components for importing relational data and analyzing file based data using NextCloud and JupyterHub. The platform ensures component interoperability, standardization of service creation processes, and infrastructure scalability. Over 200 services and five geoportals have been deployed on its basis. The proposed solution reduces barriers to interdisciplinary collaboration by standardizing data collection, storage, and processing. The architecture’s flexibility, in collaboration with the use of data models, inheritance mechanisms, and containerization, enables component reuse and supports the full lifecycle of scientific data in accordance with the principles of FAIR and open science.
В статье строится и исследуется математическая модель горно-перевального участка железной дороги, на котором для движения грузовых поездов в гору требуется подталкивание, в связи с чем он является одним из лимитирующих на Транссибирской магистрали. Участок проходит по берегу Байкала, что накладывает особые экологические требования к его эксплуатации и возможным мерам по модернизации. Модель имеет вид полуоткрытой сети массового обслуживания с делением и слиянием заявок. На основе результатов численного исследования модели определяются максимальная допустимая загрузка участка, его узкие места и зоны, подверженные наибольшему техногенному влиянию. The paper presents and analyzes a mathematical model of a mountain-pass railway section where pusher locomotives are required for uphill freight trains. This makes it one of the bottlenecks on the Trans-Siberian Railway. The section runs along the shore of Lake Baikal, which imposes strict environmental constraints on both its operation and any potential upgrades. The model is a semi-open queuing network which include fork and join nodes. It consists of three types of nodes: standard nodes simulate train movement between stations and within stations; join nodes describe the coupling of pusher locomotives to trains; fork nodes represent the uncoupling process. A marked markovian arrival process describes the arrival of transit trains of different categories. The routes of the trains and pusher locomotives are captured using multiple request types with individual route matrices. The mathematical model is implemented as a simulation model and analyzed numerically. The results obtained lead to the following conclusions. The section has a 12 % capacity reserve. However, further increases in capacity will require expanding the infrastructure at Slyudyanka–2. The station is located in close proximity to the shoreline, which significantly limits modernization options. The longest freight train stopping times are observed at Slyudyanka–1 and Slyudyanka–2 stations, necessitating an assessment of the insulation quality in these areas to minimize the ingress of contaminated wastewater into Lake Baikal. The number of pusher locomotives required to maintain normal operation under future increasing traffic demand is also estimated.
In this study, starting from the integral differential model of a DC-DC boost converter with a PI controller, we construct two mathematical representations of the problem. First, the original system is transformed into a nonlinear differentialalgebraic equation (DAE) in order to clarify the role of the algebraic constraint in the control model of the DC-DC boost converter. Next, based on linearization around an operating point, an integral-algebraic equation (IAE) is established to describe the small-signal problem. For both representations, the least squares method (LSM) is proposed as a unified tool for constructing approximate solutions. The results show that, for the nonlinear DAE form of the DC-DC boost converter with PI controller, the LSM provides highly accurate approximations when compared with the reference solution obtained by the Runge-Kutta method of orders 4 and 5 (RK45). For the IAE model of the smallsignal problem, this study demonstrates that integral-algebraic equations arise naturally in power electronics, thus broadening the range of applications for this type of equation. In addition, we present the formulation and analysis of the LSM for the IAE form of the problem under consideration. The obtained results suggest a promising research direction for power electronics models with differential-algebraic and integral-algebraic structures, emphasizing the advantages of a global optimization-based approach in handling problems whose solutions may be influenced by singular points or special operating regimes.
Exact solutions to some multidimensional generalized Monge–Ampère equations are found. These solutions are a superposition of a quadratic form of spatial variables and solutions to nonlinear ordinary differential equations generated by the Monge–Ampère equations.
We study systems of two equations with the Monge–Ampère operator, whose right-hand sides may depend on the Laplace operator and the gradients of the unknown functions. To construct exact multidimensional solutions in the case where the right-hand sides involve power or exponential functions of the unknowns, we propose a reduction method to a system of ordinary differential equations. We obtain exact multidimensional solutions expressed explicitly as a superposition of quadratic forms in spatial variables and elementary functions. A series of examples of explicit exact solutions is presented, including global positive solutions and solutions anisotropic in the spatial variables.
