HPU2 Journal of Science: Natural Sciences and Technology

An eigenvalue approach to the dynamics of supply-demand-price systems

Thanh-Huyen Pham Thi, Anh-Thang Le — Thu, 28 Aug 2025 00:00:00 +0700

This paper examines the oscillatory behavior of supply-demand-price dynamical systems using eigenvalue analysis. Unlike traditional stability assessments, our study reveals that the system does not exhibit asymptotic stability since all eigenvalues have zero real parts. This results in sustained harmonic oscillations rather than convergence to equilibrium. By formulating market dynamics through differential equations and analyzing the Jacobian matrix, we characterize the system’s long-term behavior based on its eigenvalues. Our findings provide mathematical formulations, theoretical insights, and numerical simulations that illustrate persistent price fluctuations and cyclical market behavior. The study enhances the understanding of market instability, hence emphasizing the role of linear algebra in economic dynamics and its implications for economic modeling and policy-making.

Utilizing LSTM and transformer models to analyze and predict potential career paths through student scores

Le-Hang Le — Thu, 28 Aug 2025 00:00:00 +0700

Long Short-Term Memory (LSTM) models are a type of recurrent neural network (RNN) designed to capture long-term dependencies in sequential data, and mitigate the vanishing gradient problem that limits traditional RNNs. LSTMs achieve this by using gates that control the flow of information, allowing memory maintenance over time. In contrast, Transformer models, which rely on self-attention mechanisms rather than recurrence, have revolutionized the field of natural language processing. Transformers enable parallel processing of sequences, making them more efficient and scalable for tasks like language translation and text generation. While LSTMs prove to be effective for certain sequential tasks, Transformers have generally shown greater performance due to their ability to handle longer sequences and capture complex dependencies. The advent of sophisticated machine learning techniques has revolutionized the field of predictive analytics, particularly in the realm of education. This article explores the utilization of Long Short-Term Memory (LSTM) and Transformer models to analyze and predict potential career paths based on student scores. By leveraging these advanced models, educational institutions can better understand student strengths and career suitability, ultimately leading to a more personalized career guidance.

Analysis of morphological characteristics and initial development of tea bag products from Camellia in Na Hang - Tuyen Quang

Xuan-Phong Ong , Thi-Thuong Ngo, Viet-Hong La — Thu, 28 Aug 2025 00:00:00 +0700

The Yellow camellia was first discovered in Vietnam in the early 20th century. The leaves of the camellia plant contain many phenolic compounds such as ellagitannin, taxifolin deoxyhexose, proanthocyanidin, kaempferol derivatives, apigenin derivatives, glucosyl isorhamnetin, quercetin derivatives and platphylloside. Using morphological comparison methods, we initially identified the large yellow-flowered camellia in Khau Tinh commune, Na Hang district, Tuyen Quang province as Camellia megasepala Hung T. Chang & Trin Ninh. To create tea bags from yellow camellia leaves, we have built a production process and with evaluated food hygiene and safety criteria, and as well as assessed sensory quality criteria according to TCVN 3218: 2012) is built. The evaluation results indicated that the tested formulas achieved high sensory scores, all exceeding 19.5 points.

Assessment of antibacterial activity of Spirulina platensis cultivated on camel urine medium, in vitro

Ahmed-Amrajaa Abdulrraziq, Sami-Mohammed Salih, Amani-Amrajaa Abdulrraziq — Thu, 28 Aug 2025 00:00:00 +0700

This research sheds new light on the evolution of microalgae behaviour, specifically their interaction and adaptation strategies to atypical media. The present study was conducted to evaluate the antibacterial activity of Spirulina platensis cultivated in a medium made from camel urine at concentrations of 1:1, 1:2, 1:3 and 1:4 (v:v) during periods (7, 28, and 42 day). The test was performed by the disc diffusion method, against five types of human pathogenic bacteria [Escherichia coli (ATCC 25922), Pseudomonas aeruginosa (ATCC 10145), Staphylococcus aureus (ATCC 25923), Methicillin-Resistant Staphylococcus aureus and Proteus spp]. Results showed that S. platensis grew and adapted well in most camel urine media, especially in T2 medium with biomass (5.1 g/L), chlorophyll a and b (2.4 and 1.2µg/ml), carbohydrates 40.0% (w/w), and total protein (24µg/ml). In contrast, the bacterial susceptibility test's results were unexpected. The maximum rate of pathogen-bacteria inactivation was attained in the last phase following 42 days. In detail, Significant activity was observed for S. platensis extracts cultivated in low-concentration camel urine medium (T4) against all bacterial species compared to S. platensis extracts naturally cultivated in standard medium. Staphylococcus aureus was the most sensitive species of bacteria to S. platensis extracts with an inhibition zone of 5.0 cm.

Evaluation of hybridization potential and identification of molecular markers for early recognition of f1 hybrids between Cymbidium aloifolium (L.) Sw. 1799 and Cymbidium finlaysonianum Lindl. 1833

Phuong-Thu Pham, Kim-Anh Duong Thi, Huyen-My Tran — Thu, 28 Aug 2025 00:00:00 +0700

This study focuses on evaluating the hybridization potential between two native Vietnamese orchid species, ♀ Cymbidium aloifolium (L.) Sw. 1799 and ♂ Cymbidium finlaysonianum Lindl. 1833, and identifying specific SSR molecular markers for early recognition of F1 hybrids. The hybridization process was carried out using emasculation and cross-pollination methods, followed by the development of F1 populations under in vitro conditions. After eight months, the hybrid fruits were harvested, and their seeds germinated successfully on MS medium supplemented with 3% sucrose and 5 g/L agar. A substantial number of F1 individuals were obtained, of which 250 were selected for SSR molecular analysis. Eighteen SSR markers were tested; eight primers exhibited clear amplification, and two primers (CS16 and CS7) showed distinct polymorphism between parental samples. Based on the analysis, eight representative F1 individuals were selected and further evaluated using the CS7 marker, confirming allele inheritance from both parents and thereby validating the success of hybridization. These results contribute valuable molecular tools for early-stage identification of true hybrids and provide a scientific foundation for breeding programs and conservation efforts targeting native Cymbidium germplasm in Vietnam.

Taxonomic characteristics of the Genus mucuna Adans. in Vietnam

Thi-Thuy Nguyen, Hong-Anh Le, Quynh-Chi Pham, Thanh-Hoa Mai, Minh-Tam Ha — Thu, 28 Aug 2025 00:00:00 +0700

The genus Mucuna Adans. encompasses around 105 species, predominantly inhabiting tropical and subtropical zones, with a significant number occurring in Asia. These species are mostly associated with primary and secondary forest habitats and are seldom cultivated, with M. pruriens being a notable exception. They are characterized as woody lianas, either annual or perennial-they exhibit considerable morphological variation. Mucuna species significantly contribute to forest ecosystem dynamics and offer diverse applications, including in timber production, traditional medicine, and food resources. Of these, two species are cultivated, while three are considered threatened and of conservation concern. This study provides a comprehensive account of the genus Mucuna in Vietnam, detailing its key morphological features, distribution patterns, habitat preferences, ecological functions, and practical applications. Additionally, a dichotomous identification key is proposed to aid in distinguishing the seven native species found in the flora of Vietnam.

Morphology and anatomy characterization of Smilax gaudichaudiana Kunth in Vietnam

Diem-Quynh Le, Mai-Huong Pham, Lan-Huong Do Thi — Thu, 28 Aug 2025 00:00:00 +0700

Smilax gaudichaudiana Kunt woody vine, the stem shows a clear distinction between the bark and the column, which is a unique characteristic of Smilacaceae, differing from the common features of Monocotyledoneae. A part of the rhizome of Smilax gaudichaudiana Kunth is transformed into a tuber that has medicinal properties. The leaves are glossy green, oval, egg-shaped or heart-shaped, with parallel veins and undifferentiated palisade tissue, which is suitable for dense forest conditions with low light. The roots develop strongly in the form of fibrous roots, with an anatomical structure featuring a Caspari belt that persists throughout the life of the plant. The diameter of the roots is small and hard, and the soft tissue becomes woody, helping the supporting column to be more solid. The flowers are small, gathered in umbels, unisexual, different from each other, growing in the leaf axils. The fruit is spherical, 0.8–1 cm in diameter, red-black when ripe, with 1–4 seeds.

An identification problem governed by nonlinear fractional mobile-immobile equation, Part I: Solvability

Van-Dac Nguyen, Thi-Thu Tran , Van-Tuan Tran — Thu, 28 Aug 2025 00:00:00 +0700

In this study, we address the inverse source problem of identifying a space-dependent parameter in nonlinear fractional mobile-immobile (FrM-IM) equations. The inverse problem is resolved using supplementary measurements taken at the final time, which are permitted to depend implicitly on the system’s state. This work is presented in two parts. In Part I, we first establish regularity estimates for resolvent operators associated with the linear FrM-IM equation under Dirichlet boundary conditions. Due to these estimates, we employ fixed-point arguments and local analysis on Hilbert scales to rigorously prove the existence and uniqueness of solutions to the nonlinear inverse problem. In Part II (to be addressed separately), under sufficient regularity assumptions on the final datum and the governing nonlinearities, we demonstrate that the solution derived in Part I is, in fact, a strong solution. Our analysis advances the theoretical framework for FrM-IM equations by unifying resolvent operator theory with nonlinear fixed-point methods, thereby providing a foundation for addressing inverse problems in nonlocal transport phenomena.

Final value problem for fractional reaction-subdiffusion equations

Thanh-Tuan Pham, Thi-Ngan Nguyen — Thu, 28 Aug 2025 00:00:00 +0700

We investigate the existence of a mild solution to the final value problem for a class of fractional reaction-subdiffusion nonlinear equations, where the nonlinearity may take weak values. We want to demonstrate the unique existence of a mild solution by using the Banach fixed-point theorem. In order to do this, we construct some new estimates for the resolvent function and the resolvent operator, based on the existing resolvent theory. From our point of view, the nonlinearity, which takes values in Hilbert scales, presents some technical difficulties but allows us to examine broader classes of problems, since by which it can contain a polynomial or gradient term arising from various physical circumstances.

Study of the melting of FeSi interstitial alloy films by SMM

Thu, 28 Aug 2025 00:00:00 +0700

This study utilizes the statistical moment method to investigate the effects of high temperature and pressure on the structure and melting process of Fe thin films and FeSi interstitial alloy thin films with a BCC structure. The theory is developed and validated through calculations with Fe films, then extended for application to FeSi films. When the number of layers increases to 200, corresponding to a film thickness of 70 nm, the melting temperature approaches that of the bulk material. The research results indicate that the addition of Si to FeSi films causes significant changes in melting temperature and structure, reducing the material's heat resistance.