OJS Tashkent State University of Economics
Not a member yet
554 research outputs found
Sort by
ОСОБЕННОСТИ ВНЕДРЕНИЯ ТЕХНОЛОГИИ WTTX В СЕТЯХ ТЕЛЕКОММУНИКАЦИИ
No full textВ данной статье приведен анализ архитектуры технологии WTTx (Wireless to the x), особенности развития и внедрения на сетях телекоммуникации. Технология WTTx является важным направлением развития современных беспроводных технологий, обеспечивая высокоскоростной доступ к сети широкополосного Интернета для конечных пользователей. Показан подробный анализ основных принципов построения архитектуры WTTX, особенности технологии, а также проведена оценка её эффективности и потенциала для применения в современных телекоммуникационных сетях
MODEL-BASED ASSESSMENT OF ENERGY AND RELIABILITY EFFICIENCY IN WIRELESS SENSOR NODES
No full textWireless sensor networks (WSN) are increasingly used in various control and management applications, where the efficiency of sensor nodes plays a decisive role in the overall operation of the network. This work presents a model-based assessment approach for assessing the energy efficiency and reliability of wireless sensor nodes. The proposed models analyze the relationship between transmission energy, node operation time, and network reliability under various operating conditions. Simulation-based estimates show how energy constraints and packet loss levels affect the stable operation of sensor nodes. The study compares the effectiveness of several existing energy models and provides recommendations for optimizing the behavior of nodes in energy-restricted environments
OLORA+: A HYBRID APPROACH TO LOW-RANK ADAPTATION
No full textThis paper introduces and analyzes OLoRA+, a novel parameter-efficient fine-tuning method that enhances Low-Rank Adaptation (LoRA) by combining the orthonormal initialization of OLoRA with the differential learning rate optimization of LoRA+. The study compared the performance of OLoRA+ against the standard OLoRA baseline on the TinyLlama-1.1B-Chat-v1.0 model. Using a subset of the alpaca dataset, a collection of 52K instruction-following demonstrations, with 1000 samples for training and 500 for testing, model performance was assessed using evaluation loss, BLEU, and ROUGE metrics. The obtained results show that OLoRA+ consistently outperforms the OLoRA baseline across all considered metrics. Additionally, the study reveals that OLoRA+ is effective with learning rate ratios both greater and less than one, uncovering a novel trade-off between "Refinement" and "Exploration" learning strategies. This confirms OLoRA+\u27s potential as a more versatile and powerful approach for LLM adaptation under limited resource conditions
YEYILISH DINAMIKASINI TADQIQ ETISHDA EMPERIK TENGLAMALARNI KELTIRIB CHIQARISHDA MAHSUS DASTURLARDAN FOYDALANISHNING ISTIQBOLLARI
No full textMaqolada dala tajriba sinovlaridan olingan ma‘lumotlar asosida o‘qyoysimon panjaning еyilishi uchun emperik tenglamalarni keltirib chiqarilish uchun mahsus dasturdan foydalanish va bu usul bilan tadqiqot ishlari uchun vaqtni tejab, katta hajmda еyilishini o‘rganish tadqiqotlarini olib borish hamda ishonchli natijalar olish mumkinligi kеltirilgan
TIBBIY MA’LUMOTLARNI TAHLIL QILISHNING INNOVATSION ALGORITMLARI ORQALI INSULTNI ERTA ANIQLASH
No full textUshbu maqolada ko‘p tipli ma’lumotlar — nominal va qiymatli belgilardan tashkil topgan o‘quv tanlanmalarni tahlil qilishda Fisher mezoniga asoslangan informativ belgilarning tanlanishi masalasi yoritiladi. Ma’lumotlarning turlicha tabiati tahlil jarayonini murakkablashtirishi sababli, sinflararo farqlanish va sinflar ichidagi o‘xshashlik darajasini aniqlash imkonini beruvchi funksional mezonlardan foydalanish muhim ahamiyatga ega. Shu nuqtai nazardan Fisher mezoni eng samarali yondashuvlardan biri sifatida ko‘rib chiqilib, optimal informativ belgilarning aniqlanishi bo‘yicha ikki bosqichli algoritm taklif etiladi. Birinchi bosqichda qiymatli va nominal belgilarga asoslangan optimizatsiya masalasi yordamida informativ vektorlar tanlanadi, ikkinchi bosqichda esa sinf obyektlarining o‘zaro o‘xshashlik darajasi hisoblanib, optimal kombinatsiya aniqlanadi. Taklif etilgan usul turli tipdagi ma’lumotlar fazosida samarali natijalar berishi va tahlil jarayonining aniqligini oshirishi bilan ahamiyatlidir
SELF-ADAPTING HYBRID EVOLUTIONARY ALGORITHM FOR TRANSPORT LOGISTICS OPTIMIZATION: DEVELOPMENT, STABILITY ANALYSIS AND PRACTICAL APPLICATION
No full textThis paper presents a self-adaptive hybrid evolutionary algorithm for solving transport route optimization problems in agricultural logistics. The developed algorithm combines differential evolution (DE), a genetic algorithm (GA), and variable-environment local search (VNS) with a self-adaptive parameter optimization (SaDE) mechanism. SaDE automatically adjusts key algorithm parameters based on their success in previous generations, ensuring an adaptive balance between global search and local exploitation. A comprehensive analysis of the algorithm\u27s stability was conducted on problems of varying dimensions (from 50 to 200 orders). Experimental results demonstrate a 15-18% reduction in transportation costs compared to baseline heuristics with a variation coefficient of only 1.40%, which is 3.5 times better than a classical genetic algorithm. The high stability and reproducibility of the results confirm the algorithm\u27s readiness for integration into real-world transport logistics management systems
KATTA HAJMLI MA’LUMOTLAR TIZIMIDA AXBOROT XAVFSIZLIGI RISKLARINI ANIQLASHGA ASOSLANGAN FOYDALANISHNI BOSHQARISH
No full textMazkur maqolada katta hajmli ma’lumotlar tizimlarida axborot xavfsizligini ta’minlash uchun riskka asoslangan foydalanishni boshqarish mexanizmini ishlab chiqish masalalari ko‘rib chiqilgan. Tadqiqotda foydalanishni boshqarish jarayoni va arxitekturasi tahlil qilinib, riskni asosiy baholash mezoni sifatida tanlashning maqsadga muvofiqligi asoslab berilgan. Taklif etilgan mexanizm foydalanishlarni boshqarish modeli asosida qurilgan bo‘lib, foydalanuvchi va obyekt atributlarining dinamik yangilanishi, risk qiymatini real vaqt rejimida baholash hamda foydalanuvchi vakolatlarini moslashtirish imkonini beradi. Mexanizm to‘rtta asosiy moduldan iborat: tayanch struktura moduli, riskni aniqlash moduli, foydalanuvchilarni klasterlash moduli va foydalanuvchi atributlarini yangilash hamda foydalanish siyosati moduli. Riskni hisoblashda foydalanuvchilarning katta hajmli ma’lumotlar tizimidagi xatti-harakatlari Yevklid masofasi va statistik ko‘rsatkichlar asosida baholanadi. Shuningdek, foydalanuvchilar risk qiymatlariga ko‘ra aglomeratsiyalangan iyerarxik klasterlash algoritmi yordamida tasniflanadi. Taklif etilgan yondashuv katta hajmli ma’lumotlar tizimida foydalanish huquqlarini moslashuvchan, xavfsiz va real vaqt rejimida boshqarish imkonini beradi
MA’LUMOTLARNI SHIFRLASH ALGORITMI AKSLANTIRISHLARINING ALGEBRAIK XUSUSIYATLARI TAHLILI
No full textMazkur maqolada O‘zDSt 1105:2009 algoritmida ishlatilgan akslantirishlarning algebraik xususiyatlari tahlil qilinadi. Tadqiqotda algebraik kriptotahlil usuli asosida O‘zDSt 1105:2009 algoritmi komponentlarining algebraik tenglamalar orqali ifodalanishi, ularning xususiyatlari va samaradorlik tahlili amalga oshirilgan. Shuningdek, algoritm komponentlari uchun ishlab chiqilgan tenglamalar sistemalari va ularni tahlil qilish usullari ko\u27rib chiqilgan. Bu usullar orqali algoritmning maxfiy kalitiga oid ma\u27lumotlarni aniqlash imkoniyati tahlil qilingan. Tadqiqot natijalari algebraik kriptotahlilning samaradorligini oshirishga qaratilgan tavsiyalarni o‘z ichiga oladi
MICROSOFT WORD UCHUN O‘ZBEK TILI SO‘ZLARNI IMLOVIY TAHRIRLASH PLAGININI YARATISH ALGORITMLARI VA DASTURI
No full textGlobalizatsiya va raqamli axborotlashtirish jarayonida o‘zbek tilining rivojida axborot texnologiyalari muhim ahamiyat kasb etmoqda. Kompyuter lingvistikasining tabiiy tilni qayta ishlash NLP (Natural Language Processing) yo‘nalishi til bilan bog‘liq masalalarni, jumladan, tahrir, tahlil, tarjima, elektron matnni ovozlashtirish, og‘zaki nutqni elektron matnga aylantirish, va boshqa ko‘plab jarayonlarni avtomatlashtirish uchun dasturlar yaratishga xizmat qilmoqda. Ushbu maqolada Microsoft Word muhiti uchun o‘zbek tili so‘zlarini imloviy tahrirlashga mo‘ljallangan plagin yaratish va uni amalga oshirish jarayoni ko‘rib chiqiladi. Plaginning asosiy vazifasi – o‘zbek tilidagi matnlarni avtomatik tahrirlash va xatolarni aniqlash orqali foydalanuvchilarga qulaylik yaratish. Ushbu tizim o‘zbek tilidagi matnlarni morfologik va grammatik jihatdan tahlil qilib, xatolarni aniqlaydi va ularni tuzatish imkonini beradi. Maqolada plaginni yaratish jarayoni, uning algoritmlari va Microsoft Word muhiti bilan integratsiyasi batafsil yoritiladi
ПОСТРОЕНИЕ ДИНАМИЧЕСКИХ МОДЕЛЕЙ ТЕХНОЛОГИЧЕСКИХ СИСТЕМ ОБРАБОТКИ НЕЖЕСТКИХ ОСЕСИММЕТРИЧНЫХ ДЕТАЛЕЙ
No full textВ статье рассматривается построение динамических моделей технологических систем обработки нежестких осесимметричных деталей, особенно в контексте точения на токарных станках. Основное внимание уделено математическому описанию динамических свойств технологической системы, включая взаимодействие упругой системы и процесса резания. Описание строится на основе нелинейных уравнений и передаточных функций, учитывающих влияние различных факторов, таких как упругие деформации и изменения припуска на обработку. Также обсуждаются методы идентификации модели и анализ устойчивости системы управления для минимизации погрешности формы детали. В статье показано, как выбранная модель может быть использована для оценки влияния динамических характеристик на процесс обработки, а также для разработки методов активного подавления колебаний