2949-1290 5 1 16 2026 1-102

UNK RUS 4-19 Peter the Great St.Petersburg Polytechnic University Yakovleva Alena

Saint Petersburg, Russia

57210345222 0000-0002-4822-6768 Peter the Great St.Petersburg Polytechnic University Levina Anastasia

Saint Petersburg, Russia

Regional Digital Infrastructure: Key Elements and Their Interrelations The object of this study is the digital infrastructure of the constituent entities of the Russian Federation. The subject of the study is the structural interrelations between the elements of digital infrastructure within the regional context. The methodological framework comprises a systems approach to analyzing infrastructure as a multilevel phenomenon, a comparative analysis of statistical data from the Ministry of Digital Development of the Russian Federation for 2022–2023, and a case study method for an in-depth examination of practices in three types of regions: a metropolitan region (Moscow), a digitalization leader (Tatarstan), and a typical agrarian region (Kursk Oblast). The study reveals a persistent differentiation among regions in terms of digital infrastructure development: the gap between the most and least developed entities in network capacity reaches a factor of 4.7. Three groups of systemic problems hindering effective interaction among infrastructure elements are identified: economic (the cost of laying fiber-optic communication lines in rural areas reaches RUB 2.8 million/km), technological (63% of regional information systems use foreign software), and human capital (an annual outflow of 18.7% of IT specialists from regions). It is established that sanctions pressure has accelerated import substitution (the share of domestic software in the public sector increased from 35% to 65%) but has led to delays in the implementation of infrastructure projects in 40% of regions. Practical recommendations are developed for federal authorities, regional governments, and the business community aimed at optimizing the architecture of digital infrastructure, taking into account the specific characteristics of different types of regions. An integrative model of regional digital infrastructure is proposed, encompassing structural, spatial, institutional, and technological sovereignty components. 10.57809/2026.5.1.16.1 330.47 digital infrastructure enterprise architecture regional development technological sovereignty systems analysis spatial economics import substitution human capital potential digital transformation https://technoeconomics.spbstu.ru/article/2026.16.1/ 1_yakovleva_levina.pdf

UNK RUS 20-31 Peter the Great St.Petersburg Polytechnic University Khafetulin Artur

Saint Petersburg, Russia

0000-0003-1162-7733 Peter the Great St.Petersburg Polytechnic University Gugutishvili Dayana

Saint Petersburg, Russia

Development of an Innovative Strategy for QSR Chain Expansion into Asian Markets: A Case Study of Dodo Pizza The article addresses the strategic challenges faced by Quick Service Restaurant (QSR) chains when entering highly competitive and culturally diverse Asian markets. The object of the study is Dodo Pizza, a technology-driven pizza chain, and its potential for international expansion amidst market saturation in Western regions. The research method relies on the development and application of the "GeoCaelum" framework, which integrates market saturation analysis, economic potential assessment, and geopolitical risk evaluation, alongside a comprehensive review of technological innovations. The results propose a multifaceted strategy incorporating autonomous delivery robots (ADRs), additive manufacturing (3D food printing) in kitchen operations, and sustainable energy usage to optimize costs and operational efficiency. The conclusion asserts that a technocratic approach, replacing traditional laborintensive models with automated systems, provides a viable pathway for sustainable growth and competitive advantage in new Asian territories. 10.57809/2026.5.1.16.2 339.9:005.21 international business strategy Asian markets QSR industry Dodo Pizza innovation management autonomous delivery robots additive manufacturing GeoCaelum system https://technoeconomics.spbstu.ru/article/2026.16.2/ 2_khafetulin_gugutishvili.pdf

UNK RUS 32-40 Peter the Great St.Petersburg Polytechnic University Cherepanov Saveliy

Saint Petersburg, Russia

Hybrid AI models: a combination of classical algorithms and neural networks to enhance interpretability Modern Deep Learning neural networks demonstrate high accuracy in classification and forecasting tasks, but significant limitations remain in the interpretability of the results. This creates an obstacle to application in critical areas where maximum transparency of decisions is required. In this paper, we propose the use of a hybrid approach that combines both feature extraction methods based on neural networks and classical interpreted machine learning algorithms. In the course of the work, an architecture was developed in which a neural network forms a compact representation of data, and the final decision is made by an interpreted model in the form of a decision tree or logical regression. Experiments have been conducted on open datasets, confirming that the proposed approach allows for increased interpretability while maintaining accuracy comparable to Deep Learning models. The results demonstrate the promise of hybrid architectures for areas requiring transparency and explainability of the results. 10.57809/2026.5.1.16.3 330.47 neural networks interpretability hybrid models data analytics decision-making https://technoeconomics.spbstu.ru/article/2026.16.3/ 3_cherepanov.pdf

UNK RUS 41-51 Apadana Institute of Higher Education Mehri Nahid

Shiraz, Iran

Large Language Models (LLMs) in E-Commerce Large language models (LLMs) have emerged as highly influential technologies in e-commerce, offering possibilities for complex applications in customer interaction, business efficiency, and decision-making. LLMs such as GPT-4, BERT and t5 have demonstrated high accuracy in natural language processing with transformer-based deep learning architecture and contribute to personalized recommendations, content creation and intelligent customer engagement. Existing literature has shown that LLM-based chatbots have the potential to solve more than 80% of common customer questions. Similarly, the recommendations made by LLMs have shown a significant contribution in terms of revenue growth for e-commerce businesses. Despite these potential applications, there are still major challenges in implementing LLMs in e-commerce, including privacy, cost and ethics issues. The purpose of this paper is to systematically study the application of LLMs in e-commerce, including its advantages, disadvantages and potential. With comparative case study approaches for Amazon's three e-commerce giants, Alibaba and Shopify, and by analyzing emerging trends in multifaceted AI and voice commerce, while considering the key implementation challenges, the research seeks to provide valuable insights to optimize the application of LLMs in e-commerce. 10.57809/2026.5.1.16.4 330.47 large language models e-commerce natural language processing chatbots https://technoeconomics.spbstu.ru/article/2026.16.4/ 4_mehri_nahid.pdf

UNK RUS 54-63 0000-0002-1341-2288 Peter the Great St.Petersburg Polytechnic University Frolov Konstantin

Saint Petersburg, Russia

Multicriteria Aspect of Optimal Choice Model in the Adaptive Resource Management Problem of a Manufacturing Company An approach to selecting a resource allocation option in economic systems based on a single criterion does not align with the objective reality of management. In this regard, the emphasis on multi-criteria choice, which is most relevant for a manufacturing company, requires no justification as it is more attractive in terms of reflecting the objective scenario of resource management. The aim of this study is to develop an abstract formalized model and methodological support for the theoretical-methodological apparatus that incorporates several choice criteria: flexibility in scheduling operations (operational flexibility), stability, and economic efficiency. These aspects of activity should not be viewed as static entities but rather with consideration of dynamics arising from the organization’s functioning in a competitive environment, changing principles of fiscal regulation, and the influence of natural factors. The methodological basis of the research includes multi-criteria choice theory, operations research, and adaptive control theory (Lotov and Pospelova, 2008). The paper proposes a model intended for use within a rolling planning horizon; the model incorporates a mechanism for dynamic calibration of weight coefficients based on Bayesian updating and an algorithm for constructing the Pareto front. Approaches are proposed for assessing key performance indicators related to resource allocation, delays in operational decision-making, and the enterprise’s ability to respond to unplanned disturbances. The work may be useful in the context of developing the theory of adaptive control in economic systems; the proposed provisions can serve as arguments for designing tools for intelligent decision support systems in manufacturing companies. 10.57809/2026.5.1.16.5 330.47 multi-criteria optimization adaptive resource management economic model Pareto optimality manufacturing company dynamic allocation supply chain resilience https://technoeconomics.spbstu.ru/article/2026.16.5/ 5_frolov.pdf

UNK RUS 64-74 Hubei Engineering University Liu Dongxu

Xiaogan City, Hubei Province, China

Reconstructing the Agility of the Prefabricated Building Supply Chain The prefabricated building supply chain faces challenges in multi workshop coordination, high costs, and weak resilience. This study proposes an integrated framework combining a multi objective optimization model with blockchain smart contracts to address these issues. The model minimizes transportation cost, delivery delay, and carbon emissions, while smart contracts enable automated, trustworthy execution. Case study results show transportation cost reduced by 25.9%, on time delivery increased by 17.9%, carbon emissions cut by 28.8%, and the default rate dropped from 8.5% to 2.1%. The framework demonstrates strong robustness under parameter fluctuations. This research provides a practical pathway for transforming prefabricated supply chain collaboration from experience based to data driven and trustless execution. 10.57809/2026.5.1.16.6 330.47 prefabricated building supply chain collaboration multi-workshop optimization blockchain agile supply chain https://technoeconomics.spbstu.ru/article/2026.16.6/ 5_frolov.pdf

UNK RUS 75-84 Peter the Great St.Petersburg Polytechnic University Ermochenko Semen

Saint Petersburg, Russia

An integrated approach to demand forecasting and inventory optimization in e-commerce This study investigates demand forecasting and inventory optimization in an e-commerce environment with a large assortment of products and highly variable demand. The research focuses on SKU-level demand modeling based on transactional data from an online retail store. The proposed approach combines machine learning methods with a stochastic inventory model. Demand forecasting is performed using regression-based and ensemble models with engineered temporal features, including calendar variables, lagged values, and rolling statistics. Demand uncertainty is estimated based on forecasting errors and adjusted using a robust capping procedure. The results show that the gradient boosting model provides the highest forecasting accuracy (MAE = 23.97, RMSE = 311.70). The average weekly demand across products is approximately 158 units, while demand variability differs significantly between SKUs. The application of the Newsvendor model leads to an average optimal order quantity of 236 units, which reflects the impact of demand uncertainty on safety stock formation. However, unconstrained solutions exceed the available budget by more than 14 times. To address this issue, a budget constraint is incorporated, and a proportional scaling procedure is applied. As a result, the average order size is reduced to 16 units, and the total procurement cost (119.3 thousand monetary units) satisfies the budget constraint (122.8 thousand). The achieved service level is approximately 0.94. The study demonstrates that integrating machine learning forecasting with stochastic inventory optimization provides an effective decision-support tool for e-commerce, enabling balanced consideration of demand uncertainty, service level, and financial constraints. 10.57809/2026.5.1.16.7 330.47 demand forecasting inventory optimization e-commerce machine learning Newsvendor model supply chain management stochastic optimization time series forecasting ensemble learning gradient boosting demand uncertainty safety stock https://technoeconomics.spbstu.ru/article/2026.16.7/ 7_ermochenko.pdf

UNK RUS 85-101 Peter the Great St.Petersburg Polytechnic University Kuzmenko Nikita

Saint Petersburg, Russia

Enterprise Architecture and IOT Integration in Logistics Optimization This study addresses the inefficiency of logistics systems caused by data fragmentation, lack of real-time transparency, and weak integration of business processes and digital technologies. The relevance of the study is driven by the growing demand for adaptive and sustainable supply chains in the face of global challenges and increasing digitalization. The goal of the study is to develop a unified system for integrating Enterprise Architecture (EA) and Internet of Things (IoT) technologies to optimize logistics operations. The study uses a qualitative methodology based on a systematic literature review and multivariate analysis of real-world implementations, including DHL Resilience360, Continental Tires, Union Pacific Railroad, and reference architectures based on the Internet of Things. The study follows a structured sequence: literature selection, thematic analysis, case comparison, and generalization into a generalized architectural model. The results show that integrating the Internet of Things into performance management systems enhances real-time transparency, preventive maintenance efficiency, and dynamic routing, resulting in a 20-30% improvement in efficiency. A multi-level EA-IoT architecture model is proposed, combining the levels of data collection, communication, processing, and application according to the application areas of the enterprise architecture. The research results confirm that the integration of EA-IoT provides a scalable and sustainable foundation for intelligent logistics systems and addresses the existing gaps in disparate research. 10.57809/2026.5.1.16.8 330.47 enterprise architecture Internet of Things logistics optimization supply chain management real-time tracking predictive analytics IoT sensors TOGAF framework https://technoeconomics.spbstu.ru/article/2026.16.8/ 8_kuzmenko.pdf