Islamic University of Technology

This paper presents a feed-forward neural network (NN) for hosting capacity analysis (HCA) in real-time using information from a limited number of buses, addressing the challenge of data scarcity and the need for efficient HCA with minimal input data. A limited number of buses in the distribution network is identified, referred to as pilot buses, using an approach that maximizes observability and controllability while minimizing voltage deviations at load buses to ensure network robustness. A large number of loading and electrical generation scenarios are generated using Monte Carlo simulation to train the NN with voltage information as inputs and HC as output. This trained NN can then predict the hosting capacity for distribution networks within seconds with new inputs, offering a highly efficient and practical solution for distribution utilities. The effectiveness of the proposed approach is demonstrated on the IEEE 123 bus distribution network and a rural feeder in New South Wales, Australia. Results show that the model predicts hosting capacity in 0.04 s for the IEEE 123 network and 0.18 s for the real network, with an average error of less than 1%, showcasing its accuracy and reliability. This rapid and precise prediction capability enhances distribution network management and operational efficiency for utilities. • A feed-forward neural network is proposed for quick hosting capacity analysis. • Only voltage information from pilot buses is used to train the neural network. • The method is tested on the IEEE 123 bus system and a rural feeder in Australia. • The proposed method shows promising results and requires fewer inputs.

Purpose Integration of lean manufacturing and Industry 4.0 (I4) has been a discussable topic for some years now. Integration of these two can have a synergetic effect on the system. Previous research presented conceptual frameworks for integrating lean into I4 but lacked a systematic framework. This study aims to focus on proposing a systematic framework to approach I4 from lean manufacturing. Design/methodology/approach Assembly line data collected from an industry was used to create seven simulation models using ARENA software. Two lean tools – root cause analysis and assembly line balancing have been used to identify the source of defects and unevenness in the production line, with and without automation and I4. Findings The quantified study proves the validity of the systematic framework. This framework can be applied in all industries that are thinking of implementing I4. Originality/value Previous research presented conceptual frameworks for integrating lean into I4 but lacked a systematic framework. This study focuses on proposing a systematic framework to approach I4 from lean manufacturing.

This systematic review integrates the scientific procedure SPAR-4 methodology to analyse 268 peer-reviewed human resource analytics (HRA) papers published over 14 years. A set of data from “Scopus database” was used to visualize the research landscape with the utilization of VOSviewer software. Then this study used theories, contexts, methods, and characteristics (TCCM) framework for content analysis of HRA literature based on journal indexed in the Social Sciences Citation Index (SSCI), Emerging Sciences Citation Index (ESCI), and Australian Board of Deans Council (ABDC)-indexed journals. The study evaluates performance trends in terms of publication and intellectual structure in terms of co-occurrence keywords, citation networks, bibliographic coupling the documents, and citations cited authors. The analysis reveals that growth since the early 2010s has accelerated since the pandemic. Bibliographic coupling identifies key documents and authors, whereas keyword co-occurrence analysis reveals five knowledge clusters showing the progression of technology adoption The content analysis reveals the main theories including technology organization environment (TOE), diffusion of innovation (DOI), unified theory of acceptance and usage of technology (UTAUT), adaptive structuration theory (AST), and expectancy theory (ET), which focus on organizational performance. The highlighted methods are artificial neural network (ANN), regression, structural equation modelling (SEM), and structural topic modelling approaches in the field of HRA. Research has examined HRA applications in manufacturing, small medium enterprises (SMEs), information technology (IT) firms, public and private organizations, and hospitals, whereas education sector, corporate firms, transportation, and onsite applications remain understudied. Future research directions include blockchain, metaverse, machine learning, artificial intelligence (AI), and sustainable AI integration. Researchers should focus on smart technologies to develop human resource (HR) policies, strategies, employee satisfaction, and organizational growth while addressing financial and regulatory challenges.

Automated vulnerability detection in source code is essential for modern software security, yet existing approaches suffer from high false positive rates, weak contextual understanding, and limited recall. This work presents the most comprehensive empirical evaluation to date of Small Language Model (SLM) based vulnerability detection, spanning more than twenty system configurations ranging from simple prompting baselines to advanced retrieval-augmented ensembles. Our methodology systematically examines how small LLMs (1–3B parameters) behave under different prompting strategies, static analysis integrations, ensembling techniques, and retrieval mechanisms. Through extensive ablation studies on the DiverseVul dataset, we find that many commonly assumed improvements such as complex Chain-of-Thought prompting, multi-model aggregation, or post-processing verification frequently degrade performance, largely due to overfitting and the inherent reasoning limitations of small models. In contrast, retrieval-augmented generation (RAG) consistently enhances both recall and stability, providing an average F1 gain of 4–5% by grounding model predictions in relevant vulnerability examples. Our best-performing system, a RAG-enhanced stacking ensemble with structured reasoning, achieves a 71.2% F1-score, surpassing the state-of-the-art LLaMA 3.2 model by 5.2 percentage points. our findings demonstrate that effective vulnerability detection does not arise from increased architectural complexity but from combining lightweight models with robust retrieval and context-aware reasoning. This study establishes RAG-driven small LLMs as a practical, cost-efficient, and empirically validated direction for secure software analysis. our source code at https://github.com/rafi79/RAG-Enhanced-Multi-Model-Ensemble-for-Automated-Vulnerability-Detection-Using-SLMs.

Abstract Environmental, Social, and Governance (ESG) is a well-established criterion for corporate social performance, but its financial materiality remains elusive. This study aims to unpack the nuanced and often contentious relationship between ESG and corporate financial performance, focusing not just on if ESG matters but more critically, on how, when, and for whom it becomes a significant value driver. Using a Random Forest model and SHapley Additive exPlanations (SHAP), we analyze a large panel dataset of publicly listed firms across Asia, the United States, and Europe from 2017 to 2024. The results show that firms with high ESG ratings exhibited greater resilience during the COVID-19 crisis period, confirming ESG's role as a risk mitigation factor. Moreover, our findings refine stakeholder theory by showing that satisfying stakeholder interests is not uniformly rewarded; rather, its financial benefits are conditional on both industry context and existing financial strength. These findings provide a data-driven rationale for strategic ESG investments.

This study proposes an innovative nonuniformly twisted elliptical tube heat exchanger (NTETHEX) with unequal twist pitches to address the inadequacy of uniformly twisted tube designs in double-tube heat exchanger system. A numerical investigation was conducted in the laminar regime, evaluating nine CFD cases with twist ratios ranging from 1.0 to 2.0. The optimal configuration featuring an outer tube combination of (400-800-1400) achieved the highest Nusselt number ( Nu ) and thermal performance. Further, the numerical solutions are optimized by integrating Box-Behnken Design, Response Surface Methodology, NSGA-II and a weighted factor scalarization technique. The critical design variables such as Reynolds number ( Re ), twist pitch ( P ), and surface temperature ( T s ) were analyzed to enhance the Nusselt number ( Nu ), minimizing the friction factor ( f ), and improving the PEC . Also, Quadratic polynomial regression models are developed, and their significance are evaluated through analysis of variance (ANOVA). The objective functions fitted via RSM are processed with NSGA-II to identify the set of Pareto optimal solutions. The outcomes of this multi objective optimization study compared with uniformly twisted tube and the Nu increased by 52 % and frictional coefficient is reduced by 200 % . It is evident that raising Re markedly enhances Nu . The best-performing NTETHEX configuration was found at Re = 3940.54, T s = 365 K , and P = 1.90 demonstrates the potential of nonuniformly twisted tube for highly efficient heat exchangers in critical energy applications.

As the world adjusts to the current SARS-CoV-2 epidemic, another zoonotic disease caused by a monkeypox virus has emerged across the world poses a major public health issues, demanding the use of precise and efficient diagnostic methods for detection and classification. In this paper, we offer a new deep learning framework that combines ensemble modeling with mechanisms of attention to improve monkeypox disease detection and classification. We extracted discriminative features from Monkeypox Skin Lesion Dataset version 2.0 using ten pre-trained convolutional neural network models enhanced with attention mechanisms. Later we combined the best models into ensemble model for further improvement of model performance. We proposed a majority voting ensemble model of ResNet50V2, DenseNet121, DenseNet201 where each standalone model is modified with Convolutional Block Attention Mechanism. Our approach produced remarkable results with an accuracy of 94.43% for multiclass and 96.95% for binary classification which proving our proposed model’s robustness and effectiveness in identifying monkeypox skin lesions. We used explainable artificial intelligence approaches such as Grad-CAM, SHAP, and LIME to ensure transparency and interpretability. These methods provided insights into the model’s decision-making process by emphasizing significant locations and features that contributed to the predictions. This study not only advances the field of automated monkeypox diagnosis, but it also emphasizes the need of using attention processes and ensemble strategy to improve classification performance while preserving model interpretability.