Spectrum of Engineering Sciences

PARAMETRIC OPTIMIZATION OF INCOMPRESSIBLE NAVIER–STOKES FLOW USING RESPONSE SURFACE METHODOLOGY

Ghulam Murtaza, Haider Ali — Mon, 06 Apr 2026 00:00:00 +0500

Optimization in incompressible fluid flow is a vital activity towards maximizing the aerodynamic performance and energy efficiency in a rich field of engineering systems, such as aerodynamics, HVAC networks and pipeline transport. The current paper explores the parametric optimization of incompressible Navier Stokes flow by using Response Surface Methodology (RSM). The main aim was to analyze the effect of the important design parameters as well as to determine the optimum designs that reduce the drag coefficient, the intensity of turbulence and pressure loss in the flow domain. The benchmark flow behavior and performance measures were first determined by making a baseline configuration. Response-surface models were then developed to manifest the correlation between input variables, which are the angle of the wedge, the coefficient of suction, and the concentration of nanoparticles, and the responses of the system. The optimization process produced an adjusted design which significantly enhanced the system performance compared to the baseline: the drag coefficient was reduced by about 23%, the turbulence intensity was also reduced by almost 16%, and the pressure loss was also minimized by around 14. The results show an increase in the flow stability, less energy dissipation and increased fluid transport efficiency. The formulated response-surface models had a high level of predictive power and coefficient of determination was above 0.95 in all the assessed responses. The interaction effects of the design variables between the contour plots and response-surface visualizations through graphical analyses helped to confirm the correlation. The findings prove the fact that RSM is an effective analysis tool used to investigate the complicated fluid mechanics systems and to determine the best design parameters that can be used in order to enhance the aerodynamic performance.

A TRUST-BASED ENSEMBLE MACHINE LEARNING FRAMEWORK FOR INTRUSION DETECTION IN MEDICAL INTERNET OF THINGS ENVIRONMENTS

Muhammad Bilal Abid — Thu, 02 Apr 2026 00:00:00 +0500

The high rate of Internet of Medical Things (IoMT) devices spread in clinical settings has provided critical attack surfaces that cannot be countered by conventional security measures. This paper suggests a trust-based collective machine learning framework of intrusion detecting in healthcare internet of things networks. The framework uses Mutual Information (MI) to generate dimensionality reduction by using 45 input features to generate 34 discriminative features, and then an ensemble classifier is used with hard-voting, comprising K-Nearest Neighbors (KNN), Random Forest (RF), and Support Vector Machine (SVM). The suggested model is tested on the WUSTL-EHMS-2020 data in a real-world healthcare monitoring benchmark with 16,318 network flow and patient biometric records and benchmarked with three individual baseline classifiers. The ensemble had a 95 percent accuracy, specificity of 0.99, and sensitivity (recall) of 0.65 on the minority attack class and AUC of 0.82, which is better than all the individual baselines. The training set was used only and Synthetic Minority Over-Sampling Technique (SMOTE) was used to reduce class imbalance. The findings indicate that the framework in question is capable of making reliable differentiation between malicious and regular data related to network operation, hence justifying the implementation of reliable, secure IoMT systems in mission-critical healthcare settings.

Keywords : Medical Internet of Things (IoMT); Intrusion Detection; Ensemble Learning; Mutual Information; SMOTE; WUSTL-EHMS; Trust-Based Security.

PREDICTING SCHEDULING DELAYS IN CLOUD COMPUTING SYSTEMS USING MACHINE LEARNING

Mon, 06 Apr 2026 00:00:00 +0500

Properly allocating and scheduling resources in cloud computing systems helps achieve both performance and cost goals. Poor scheduling can cause significant disruptions, which affect throughput, utilization of resources, and customer satisfaction. This study introduces a machine learning framework to predict scheduling delays for cloud instances.To predict scheduling delays for cloud instances, a complete dataset includes all requests and three types of supervised learning models (Random Forest, XGBoost, and Logistic Regression) have been evaluated. The dataset underwent many pre-processing steps, such as the elimination of leaks and the development of new features as well as the establishment of a classification target to identify which instance types had high-delays or low-delays when scheduled.

Our results demonstrate that the ensemble-based models (Random Forest and XGBoost) performed better than the linear models because XGBoost correctly predicted scheduling arrangements 74.5% of the time, while also producing reasonable results in cases of class-imbalance. The combination of feature importance analysis and SHAP interpretation indicates that the total requested resource demand for CPUs, memory limit, and instance termination time are significant contributors to delays in scheduling.

As such, the approach proposed in this paper provides cloud service providers with the means to efficiently manage scheduling delays and thereby enhance the management of resources. Overall, this research illustrates that machine learning can accurately predict scheduling outcomes for cloud computing systems, thus contributing to the transition towards a more resource-efficient cloud computing environment.

DISEASE CLASSIFICATION USING LOGISTIC REGRESSION AND MACHINE LEARNING TECHNIQUES

Azaz Ali Shah, Dr Arzoo kanwal, Amir Mushtaq, Syeda Maryam Siddiqa, Aneeza Nawaz — Mon, 06 Apr 2026 00:00:00 +0500

Accurate and early disease classification plays a critical role in improving clinical decision-making and reducing mortality associated with cardiovascular disorders. The increasing availability of medical datasets and computational tools has enabled the development of robust predictive models for disease diagnosis using statistical and machine learning approaches. A comprehensive classification framework was developed using Logistic Regression and advanced machine learning techniques for heart disease prediction based on 303 patient observations and 13 clinical predictors. The analytical framework included descriptive statistics, correlation analysis, predictor ranking, logistic regression coefficient estimation, and comparative machine learning evaluation. Multiple classification algorithms, including Random Forest, Support Vector Machine, K-Nearest Neighbors, Gradient Boosting, Decision Tree, and Logistic Regression, were evaluated using performance metrics such as accuracy, precision, recall, F1-score, and ROC-AUC. Among all models, Random Forest demonstrated the highest predictive performance, achieving an accuracy of 83.6% and ROC-AUC of 0.904, while Logistic Regression showed excellent interpretability and the highest cross-validation stability. Significant predictors included chest pain type, maximum heart rate, exercise-induced angina, oldpeak, and vessel count. The results highlight that integrating statistical inference with machine learning substantially enhances disease classification accuracy and supports reliable clinical risk assessment systems.

EFFECT OF SODIUM HYDROXIDE (NAOH) MOLARITY ON WORKABILITY AND 28 DAY COMPRESSIVE STRENGTH OF SILICA FUME AND COAL FLY ASH BASED GEOPOLYMER CONCRETE

Humayoon, Manthar Ali Keerio, Narain Das Bheel, Sadam Hussain Jakhrani — Mon, 06 Apr 2026 00:00:00 +0500

The production of ordinary Portland cement (OPC) is one of the main causes linked with carbon dioxide (CO₂) emissions which results in climate instability and environmental degradation, motivating exploration of industrial by products as sustainable supplementary Cementous materials (SCMs). This research focuses on experimental assessment of the geopolymer concrete (GPC) produced with silica fume and coal fly ash with alkaline solutions of NaOH and Na₂SiO₃. This study evaluates the effect of sodium hydroxide (NaOH) molarity and AA/B on the workability and 28 day compressive strength of geopolymer concrete made with silica fume and coal fly ash. Eighteen mix designs were prepared with three different NaOH molarities (10M, 12M, and 14M), Na₂SiO₃/NaOH ratios (1.5, 2.0, and 2.5), and alkaline activator-to-binder ratios (0.50 and 0.67). Three standard 100 mm cubes were cast for each mix, producing a total of 54 specimens. All specimens were heat cured at 90°C and then stored under ambient laboratory conditions until testing at 28 days. Workability was measured by slump, and compressive strength was determined by dividing the maximum failure load by cross sectional area of the cube. The results revealed GPC workability decreased with higher NaOH molarities, slump values averaged of 84.17 mm at 10M, 69.00 mm at 12M, and 54.00 mm at 14M. The 28 day compressive strength followed non trend, it increased from 21.92 MPa at 10M to 24.41 MPa at 12M before declining slightly to 23.63 MPa at 14M series. The highest strength mix was GPC-14, which combined 12M NaOH, Na₂SiO₃/NaOH = 2.0, and an activator-to-binder ratio of 0.50 and achieved 32.62 MPa. Finally, the identifies optimum mix demonstrates high potential for environment friendly construction. By substituting OPC with industrial by products this research contributes to the mitigation of global CO₂ emissions and offers a sustainable solution for the modern construction industry.

AI-DEVGUARD: AN ARTIFICIAL INTELLIGENCE FRAMEWORK FOR AUTOMATING AND SECURING THE SOFTWARE DEVELOPMENT LIFE CYCLE

Sherbano Saleem, Dr. Kashif Laeeq, Dr. Muhammad Asad Abbasi — Tue, 07 Apr 2026 00:00:00 +0500

The innovation of Artificial Intelligence (AI) has continued to positively impact numerous fields, including software development. In this research, we present the AI-DevGuard, an AI powered architecture designed to automate and protect subdivisions of the software development life cycle (SDLC) using intelligent tools and techniques. Using AI-DevGuard, software development teams will be able to automate the detection of bugs, improve the precision of the software, construct the software using an automated process, and improve the protection of the software through the early detection of bugs. This research will perform an in-depth analysis of the potential of AI-DevGuard in reducing the effect of human negligence, facilitating the shortening of time frames, and increasing the security in the development process. Here, we fully describe the unique plug and play design of AI-DevGuard and its compatibility to leading development tools that allows it to seamlessly integrate into the existing development tools of today’s DevOps.

A CONTEXTUALIZED FRAMEWORK FOR LOW-CARBON BUILDINGS IN DEVELOPING COUNTRIES

Abdul Jabbar, Ruhal Pervez Memon, Saira Sidhu, Qasim Raza — Wed, 08 Apr 2026 00:00:00 +0500

This study investigates the current state of low-carbon building (LCB) practices in Pakistan's construction sector and develops a contextualized implementation framework to address identified barriers. Employing a mixed-methods research design, the study integrates quantitative surveys of 153 construction professionals, qualitative semi-structured interviews with 10 industry experts, and a comprehensive literature review. Data was analyzed using descriptive statistics, thematic analysis, and comparative methods. Findings reveal that LCB adoption in Pakistan remains critically low, evident in only 5% of surveyed projects. Four primary barriers were identified and ranked: absence of regulatory frameworks and financial incentives (mean score 4.75/5), high upfront costs (4.58/5), limited awareness and technical knowledge (4.35/5), and restricted availability of sustainable materials (4.18/5). Technical analysis further demonstrates that passive design strategies can achieve 40–70% energy demand reduction, while renewable energy integration offers substantial operational carbon savings. The study contributes original empirical data on LCB implementation barriers specific to Pakistan and presents the first contextualized framework—the Three-Pillar Low-Carbon Building Design (LCBD) Framework—tailored to Pakistan's unique socio-economic, regulatory, and climatic conditions. The proposed framework provides actionable guidance for policymakers and industry professionals, with recommendations including mandatory building energy codes, targeted financial incentives, and comprehensive capacity-building programs. This research offers both theoretical contributions to sustainable construction literature and practical pathways for accelerating low-carbon transitions in developing country contexts

COMPLEX INTERVAL VALUED PYTHAGOREAN FUZZY ACZEL ALSINA AGGREGATION OPERATORS WITH APPLICATIONS

Tehreem Bibi, Ziad Khan, Rashid Jan, Shahid Iqbal, Fawad Hussain, Qaisar Khan — Wed, 08 Apr 2026 00:00:00 +0500

The common mathematical tool for combining several inputs into a single, unique result is an aggregation operator. The study introduces various aggregation operator (AOs) designed for complex interval valued Pythagorean fuzzy information. The complex interval valued Pythagorean fuzzy sets (CIVPyFS) which was created lately, proves to be a useful tool for expressing obscurity and ambiguities. The complex interval valued Pythagorean fuzzy sets have a wide range of applications in routine decision-making procedures because of their improved ability to handle uncertain circumstances compared to other fuzzy set theories. In this article, novel AOs are developed considering the advantages of the CIVPyFS to handle the multi-criteria decision-making challenges. The new AOs consider the relations between two input arguments. To improve the adaptability of the new AOs, this article incorporates the Aczel-Alsina (AA) operations. This study proposes the CIVPyF Aczel-Alsina Heronian mean (CIVPyFAAHM) operator, CIVPyF Aczel-Alsina geometric Heronian mean (CIVPyFAAGHM) operator which combines the Aczel-Alsina operational rules with the and Heronian mean/geometric Heronian mean operators. Various properties of the AOs are investigated. Further, weighted form these AOs are introduced. Then, we set up the MCDM technique using the two AOs that are suggested to solve MCDM problems under CIVPyFS environment. We next illustrate the efficacy and suitability of the predicted approach with a numerical example and compare it with other relevant MCDM strategies presently in existence in the CIVPyF information.

AI-DRIVEN SELF-REFLECTIVE MECHANISMS FOR GENERATIVE AGENTS: AUTONOMOUS PROMPT REVISION AND OPTIMIZATION

Saba Yousha, Engr. Maroof Ahmed, Muhammad Amin Gilal, Abdullah Maitlo — Thu, 09 Apr 2026 00:00:00 +0500

The speedy development of large language models (LLMs) has changed the paradigm of AI optimization to focus on fine-tuning based on weights to one based on context. This paper discusses self-reflective processes, driven by AI, to allow generative agents to automatically revise and optimize prompts without human intervention or any adjustment of parameters. We combine current developments in Agentic Context Engineering (ACE), self-reflective systems like Reflexion and new prompt optimization systems like ZERA, GreenTEA, and IROTE. Through our analysis, we have found that self-reflective architectures (including generator, reflector and curator components) always perform better than the static prompting techniques in reasoning, code generation and domain specific tasks. There is empirical evidence of 1017 percent improvement in performance and a reduction in latency of adaptation up to 87 percent. We determine three central design concepts of successful implementations namely structured feedback generation, incremental context evolution and multi-criteria evaluation. This paper ends by mentioning limitations such as computational overhead, hallucination risks, and verification errors that explain about 70 percent of reasoning failures and giving future directions of robust and verifiable self-reflection by autonomous AI systems

DEEP LEARNING-BASED ANOMALY DETECTION FOR RELIABLE INDUSTRIAL IOT MAINTENANCE

Sana Feroze, Andlib Akhtar, Aftab Rafique, Ashfaq Ahmad, Faheem Nazir — Tue, 07 Apr 2026 00:00:00 +0500

This paper introduces an AI-based anomaly detection system in Industrial IoT (IIoT) networks that can help improve predictive maintenance in intelligent manufacturing settings. The suggested model combines LSTM- autoencoders with attention systems to analyze time-series sensor data of industrial machinery, such as CNC machines, robotic arms, and PLC systems. The findings show that the accuracy of detection is very high (97 percent) and the F1-score is high (95 percent), which is much higher than the traditional rule-based systems (76 percent). The structure reduces the level of false positive (18 to 7) which increases reliability and minimizes the number of unnecessary maintenance measures.

The AI-centered model which was introduced leads to reduced downtimes by 30 percent in comparison to preventive maintenance strategies which usually only lead to improvement by 15 percent. Also, the system saves about 25 percent in costs through optimization of maintenance schedules and minimization of unforeseen failures. The edge-based processing achieves a 40 percent better real-time response through a reduction of 150 ms in the detection latency. The framework also has a high level of scalability and can support a performance efficiency of more than 89% when implemented in large scale industrial environments and 200 connected devices.

All in all, the research indicates that AI-based anomaly detection can help to enhance the efficiency of operations, the reliability of the system, and the cost-effectiveness of IIoT networks. The findings support the idea of using smart, data-driven maintenance strategies in Industry 4.0 and the need of the expanded computational efficiency and interpretability of the model in the more general industrial implementation.

LEVERAGING DEEP LEARNING AND MACHINE LEARNING IN IOT NETWORKS: ARTIFICIAL INTELLIGENCE AND CLOUD-BASED FRAMEWORKS FOR SMART AUTOMATION

Waqas Ahmad Khan, Sana Feroze, Andlib Akhtar, Ashfaq Ahmad — Tue, 07 Apr 2026 00:00:00 +0500

The article sheds light on the ways of integrating machine learning and deep learning algorithms in Internet of Things (IoT) networks with the help of cloud-computing structures and enable the powering of intelligent and efficient smart automation systems. The paper has compared the performance of various artificial intelligence systems including the random forest, XGBoost, and Neural Network in processing the IoT big data. The findings show that deep learning models and in particular the Neural Networks are the best predictors with high accuracy of 93 percent that exceeds the traditional machine learning. The paper also indicates that AI-based IoT systems would significantly enhance efficiency and performance of operations of a system would be enhanced 65 percent in conventional systems to 89 percent with the introduction of the system.

Cloud-based processing is among the critical enablers, and it increases the capacity of data processing to 90 percent and processing speed, which was 92 times more excellent compared to the performance of local processing environments. The results also focus on high accuracy of automation in all the areas of application of the industrial system, medical, agriculture and energy management with the level of performance exceeding 90 percent. Security models based on machine learning can also improve the threat detection rate to 91 percent and minimize the number of false alarms and thereby enable IoT networks to be more reliable.

The paper succeeds in concluding that the true intersection of artificial intelligence, IoT, and cloud computing presents an effective and scalable framework of intelligent automation. However, despite the problems related to the complexity of the computations and data dependency, the proposed solution demonstrates that the possibility of transforming the conventional IoT systems into intelligent, flexible, and secure ecosystems is immense. The research will result in the development of the AI-based IoT architectures of the future, and will also offer the practical recommendations on the way to make the smart environment more efficient, scalable, and decision-making.

EXPLAINABLE EDGE AI FOR EARLY DETECTION OF PLANT DISEASES IN SMART AGRICULTURE USING LIGHTWEIGHT MOBILENETV2 WITH GRADIENT SALIENCY

Thu, 09 Apr 2026 00:00:00 +0500

Plant diseases are a significant global threat to food security, resulting in annual crop losses that account for 20-40% of total agricultural products. Early and accurate plant disease detection is of primary importance for timely intervention and sustainable agricultural practices. In this paper, a new explainable edge artificial intelligence (XAI-Edge) framework for early plant disease detection in smart agriculture environments is proposed. The proposed system utilizes a light-weight MobileNetV2 architecture, where the network's width multiplier "alpha" is set to 0.75, and the model is trained on the New Plant Diseases Dataset, comprising 87,867 images from 38 plant disease classes. The proposed system focuses on a specific set of 10 plant disease classes, comprising 19,189 training images, covering early and late stages of plant diseases, including tomato, potato, corn, and apple crops. The proposed system achieves outstanding classification performance, yielding 98.54% accuracy, 98.54% F1-score, 98.56% precision, and 98.54% recall on the test set. A two-phase transfer learning approach, where the final 30 layers of the network are fine-tuned, along with a lower learning rate of 1e-5, is utilized for optimal performance. The proposed system utilizes gradient-based saliency mapping for model explainability, allowing farmers to visualize the regions of plant diseases.

COMPARATIVE ANALYSIS OF EXPLAINABLE AI TECHNIQUES FOR ENHANCED DECISION SUPPORT SYSTEMS

Thu, 09 Apr 2026 00:00:00 +0500

The rapid integration of artificial intelligence (AI) into decision support systems (DSS) has raised concerns about the transparency and interpretability of complex machine learning models. To improve the interpretability and the reliability of AI-driven decision-making, the current paper assesses the popular explainable artificial intelligence (XAI) algorithms, including LIME, SHAP, feature importance algorithm, and rule-based algorithms Experiments on benchmark datasets are used to compare these methods in regards to the explanation accuracy, consistency, computational efficiency and user interpretability. The results indicate that the combination of several XAI techniques can enhance the decision support system greatly by raising the level of transparency, user confidence and quality of decisions. SHAP based methodologies are more consistent and can be interpreted globally whereas LIME has local explanations that are able to be flexible and efficient. These improvements allow making more informed and correct decisions regarding such critical areas as healthcare and finance. The suggested research will contribute a systematic review method and practical expertise on how to select the appropriate XAI techniques and, therefore, enhance the development of a more transparent, credible and enhanced system of decision support.

APPLICATION OF SIX SIGMA DMAIC METHODOLOGY FOR REDUCING PRE-ANALYTICAL AND POST-ANALYTICAL ERRORS IN CLINICAL LABORATORY TESTING

Thu, 09 Apr 2026 00:00:00 +0500

Laboratory scientists are required to guarantee precision and accuracy in the results of diagnostic tests, as this directly impacts patient satisfaction. The Total Testing Process (TTP) of medical laboratories experiences errors that are broadly classified into three types. The errors studied in this paper are pre-analytical and post-analytical errors, and the DMAIC (Define, Measure, Analyse, Improve, and Control) methodology is applied to reduce these types of errors. Pre-analytical errors are those that happen before the testing of the sample. The majority of errors happen in this testing process. Before the introduction of the improvements, the average percentage of pre-analytical errors was recorded as 16.12%. The average percentage of post-analytical errors, which happen after the testing of the sample but before the results are delivered, was recorded as 1.01% before improvements. The results show that there are significant reductions in both stages of errors, i.e., the rate of pre-analytical errors was reduced to 5.38%, and the rate of post-analytical errors was reduced to 0.37%. The results prove that the DMAIC methodology can be applied to improve quality and reduce errors in medical laboratory testing

A GENETIC ALGORITHM-BASED APPROACH FOR DETECTING INJECTION VULNERABILITIES IN APIs

Talia Shah, Umm E Rubab, Mansoor Qadir, Sadeeq Jan — Fri, 10 Apr 2026 00:00:00 +0500

APIs play an essential role in software development in the modern world, enabling seamless communication between various applications. However, the growing trend in API vulnerabilities, particulalty related to injection raises serious security concerns. This study addresses an important gap in robust testing techniques for identifying and mitigating injection vulnerabilities in RESTful APIs. Existing automated tools have limitations, such as false positives and a lack of accuracy, that require improvement in testing methods. To addresses such security challenges, a new automated test case generation tool based on a Genetic Algorithm (GA) is presented in this study with the aim to improve the precision and accuracy of detecting injection vulnerabilities. Injection attacks, ranked eighth in the list of OWASP API Security Top 10, exploit data to manipulate interpreters, posing a huge threat to web services. Our proposed technique uses GAs that can optimize such complex problems at the highest level to provide useful test cases for maximum coverage and detect injection vulnerabilities sufficiently. The paper begins with a detailed analysis of existing approaches to API security testing and identifying vulnerabilities that are especially related to injection vulnerabilities. A new GA-based algorithm that is specially created to detect injection flaws is conceptualized after a thorough evaluation of the existing tools. The development and testing stages are aimed at ensuring reliability and efficiency, with a specific hardware-software setup being used, the performance of the tool being compared to the existing solutions.

The desired results would be to prove that the tool is highly accurate and effective when it comes to the successful identification of injection vulnerabilities. The study aims to lessen the drawbacks associated with manual testing, while enhancing the quality of testcases, and addressing resource constraints. The proposed tool will provide a proactive protection against each of the injection threats, and the API security will continue to improve. The research continues, and as we proceed, the implementation and assessment of the GA-based tool will be addressed, which will give the developers and testers valuable information to guarantee the security and integrity of web application APIs

EFFECTIVENESS OF INTRADIALYTIC EXERCISES ON HEMODIALYSIS OUTCOMES: A META-ANALYSIS

Husba Riaz, Fareha, Aqsa Mughal — Fri, 10 Apr 2026 00:00:00 +0500

Introduction:

Chronic kidney disease (CKD) has emerged as a prominent global health concern with prevalence rate ranging from 9.1% to 13.4% worldwide and 12.5% to 31.5% in Pakistan. Exercise in Hemodialysis (HD) has made significant progress in the last couple of years including resistance exercises (RE), aerobic exercises (AE) and combination exercise (CE) but still numerous questions are still unanswered.

Objective:

To determine the effectiveness of Intradialytic exercises (IDE) on HD outcomes.

Methods:

This meta-analysis was done as per PRISMA and PEDro guidelines. Data based searching was done after extraction of literature through PubMed, Embrace, and Cochrane. The methodological quality of the eligible studies was assessed. Outcome measures include Dialysis adequacy (KT/V) with a reliability of 0.98, Body mass index (BMI) with a reliability of 0.95 and HD duration

Result:

The meta-analysis of six Randomized controlled trials (RCTs) involving 293 participants showed that IDE have a little impact on BMI with SMD of 0.0691, 95% Confidence Interval [-0.1602; 0.2984], and a p-value of 0.59 and HD duration with SMD of -0.1673, 95% Confidence Interval [-0.3968; 0.0622] and a p-value of 0.1532. Meanwhile, for KT/V, IDE show a statistically significant positive effect on KT/V with SMD of 0.6458, 95% confidence interval [- 0.0264; 1.3179] with P Value 0.05, suggesting improved HD outcomes.

Conclusion:

Intradialytic exercise appears to improve KT/V; however, its effects on other clinical outcomes, including BMI and HD, remain inconclusive. These findings are limited by heterogeneity and small sample sizes, highlighting the need for larger, high-quality randomized controlled trials

AUTOMATED CUSTOMER ACCOUNT OPENING USING AGENTIC AI

Dr. Noman Hasany, Arshan Nasir, Dr. Khalid Rasheed — Fri, 10 Apr 2026 00:00:00 +0500

It takes a lot of paperwork and a physical visit to the branches to open a bank account. These documents are examined and validated by people after submission, which causes an excruciatingly drawn-out procedure. Digital banking has made an effort to provide answers over the years, but many businesses continue to rely on manual identity approvals and validation submission processes. Furthermore, these procedures lack robotic technology and intelligent automation, which restricts accessibility and ease. In order to automate every step of the process—from user interaction and identity verification to document authentication and account creation—this study presents an AI-based approach. The primary goal is to create a smart AI agent that enables consumers to safely open accounts without ever leaving their home. Uneducated or physically handicapped users can also benefit from a voice-based interface. Customers typically use traditional techniques for 180 minutes, but the suggested Chabot-based system cuts this time down to 30 minutes.

EXPLORING THE IMPLICATIONS OF CLIMATE CHANGE ON WATER RESOURCES: DEVELOPING EFFECTIVE MANAGEMENT STRATEGIES

Aisha Mughal, Muzafar Ali, Adnan Ahmed — Fri, 10 Apr 2026 00:00:00 +0500

Climate change profoundly threatens Pakistan's Indus Basin water resources, serving 240 million people across 1.2 million km² where agriculture consumes 90% of supplies amid 81 MAF deficits projected by 2025. This mixed-methods study integrates SWAT+ hydrological modeling (R²=0.87), GIS vulnerability mapping (WRI=0.48 basin-wide), CGE macroeconomic simulations ($120B annual GDP losses by 2040), and Delphi consensus from 25 experts across 7 physiographic zones. Findings reveal 22-38% runoff declines by 2100 (RCP4.5/8.5), -2.1 m/yr aquifer depletion, 28% salinization, and institutional fragmentation delaying releases 15 days/season. Rechna Doab (WRI=0.41) and Sindh delta (WRI=0.32) emerge as critical hotspots where tailenders suffer 40% supply inequities.

Pilots validate scalable solutions: managed aquifer recharge (+28% retention), drip retrofits (87% efficiency), karez rehabilitation (+22% supply). Nine interventions GIS zoning, blockchain metering, PKR 250B PPP dams—promise 3.2x ROI, lifting WRI to 0.85 and securing $5T blue economy by 2050. This triangulated framework (87% quant-qual concordance) redefines Indus resilience from scarcity narrative to efficiency arbitrage, positioning Pakistan as South Asia's water stewardship leader

CHALLENGES, OPPORTUNITIES, AND CIRCULAR ECONOMY PATHWAYS FOR CONCRETE AND CONSTRUCTION WASTE MANAGEMENT: EVIDENCE FROM URBAN BUILDING PROJECTS IN KARACHI, PAKISTAN

Abdul Jabbar, Mohsin Ali, Qasim Raza, Abdul Rehman, Sartaj Ul Nabi — Fri, 10 Apr 2026 00:00:00 +0500

The rapid expansion of urban construction activities in Karachi has led to a significant increase in construction and demolition waste, creating environmental, economic, and resource management challenges. In addition to conventional concrete debris, large volumes of secondary waste streams such as sewage sludge, plastic waste, waste glass, and tyre rubber remain underutilized despite their potential for value recovery. This study explores the challenges, opportunities, and circular economy pathways for integrated waste management in urban building projects in Karachi. Key challenges identified include the absence of effective waste segregation systems, limited recycling infrastructure, weak regulatory enforcement, and low market acceptance of recycled construction materials. However, these waste streams offer significant opportunities when incorporated into construction materials. Recycled concrete aggregates can replace natural aggregates, while processed sewage sludge ash can act as a supplementary cementitious material. Similarly, plastic waste, waste glass, and tyre rubber can be utilized to enhance concrete properties such as durability, thermal insulation, and crack resistance, contributing to sustainable material innovation.

The study proposes a circular economy framework that emphasizes waste-to-resource conversion, lifecycle extension, and reduction of virgin material consumption. By integrating multiple waste streams into construction practices, the approach not only minimizes landfill dependency but also reduces carbon emissions and promotes cost efficiency. The findings suggest that adopting such integrated circular strategies in Karachi requires coordinated policy support, technological advancement, stakeholder engagement, and increased awareness. This transition has the potential to transform the construction sector into a more sustainable and resource-efficient system while addressing the city’s growing waste management crisis.

DESIGN OF AN INTELLIGENT IOT-DRIVEN REAL-TIME CONTINUOUS REMOTE PATIENT CARE FRAMEWORK USING MACHINE LEARNING

Fri, 10 Apr 2026 00:00:00 +0500

The combination of Artificial Intelligence (AI) and the Internet of Things (IoT) has significantly transformed modern healthcare, particularly in the areas of patient monitoring and disease management. This research focuses on designing and improving an AI-enabled remote health monitoring system that utilizes IoT devices along with machine learning techniques. The primary aim is to enhance diagnostic accuracy, enable real-time data analysis, ensure data security, and achieve seamless system integration to improve overall patient care. The study also addresses key challenges such as maintaining data reliability, protecting patient privacy, and managing computational limitations. To support continuous health monitoring, various machine learning algorithms including Random Forest, Support Vector Machine (SVM), and K-Nearest Neighbors (KNN) are applied to analyze important health indicators such as heart rate, blood pressure, and ECG signals. The experimental results show that the Random Forest algorithm performs better than the other models in terms of accuracy, precision, and recall. This highlights its effectiveness for real-time healthcare applications. Overall, this research demonstrates the potential of AI-driven health monitoring systems and provides a strong foundation for developing more personalized, efficient and affordable healthcare solutions in the future.

ADVANCING LOW-CARBON AND ENVIRONMENTALLY SUSTAINABLE CONSTRUCTION MATERIALS: DESIGN AND PRACTICAL IMPLEMENTATION OF ECO-FRIENDLY CONCRETE WITH REDUCED WATER AND CEMENT CONTENT.

Omar J. Alkhatib, Syed Sajeel Ahmed, Ali Ajwad, Salman Jafar, Muhammad Riaz — Sat, 11 Apr 2026 00:00:00 +0500

The construction industry is a major contributor to global carbon emissions and resource depletion, primarily due to the extensive use of cement and water in conventional concrete production. This study presents a comprehensive approach toward the development and practical implementation of low-carbon, environmentally sustainable concrete with significantly reduced cement and water content. A systematic mix design methodology is proposed, integrating optimized particle packing, supplementary cementitious materials (SCMs), and advanced chemical admixtures to achieve enhanced performance while minimizing environmental impact. Industrial by-products such as fly ash, ground granulated blast furnace slag (GGBS), and limestone filler are incorporated as partial replacements for Portland cement to reduce clinker consumption and associated CO₂ emissions. The proposed eco-friendly concrete mixtures are evaluated through a combination of experimental and analytical techniques to assess their mechanical, durability, and sustainability performance. Key parameters, including compressive strength, workability, permeability, and resistance to chloride penetration, are systematically analyzed. The results demonstrate that the optimized mixes achieve comparable or improved performance relative to conventional concrete, despite reductions in binder and water content. The role of superplasticizers in maintaining workability at low water-to-binder ratios is also critically examined, highlighting their importance in sustainable mix design. Furthermore, a quantitative environmental impact assessment is conducted to estimate reductions in carbon footprint, water consumption, and material usage. The findings indicate that the proposed concrete design can achieve substantial reductions in CO₂ emissions and water demand without compromising structural integrity or durability. Practical implementation aspects, including scalability, cost-effectiveness, and compatibility with existing construction practices, are also discussed to facilitate real-world adoption. This study contributes to the advancement of sustainable construction materials by providing a robust framework for designing eco-friendly concrete with reduced environmental impact. The proposed approach supports global efforts toward low-carbon infrastructure development and offers a viable pathway for transitioning to greener construction practices in both developing and developed regions.

STUDY OF PHYSICAL AND ANTIMICROBIAL PROPERTIES OF BIOGENIC SILVER NANOPARTICLES SYNTHESIZED VIA GREEN ROUTE

Sat, 11 Apr 2026 00:00:00 +0500

Health is the primary concern of human beings. Apart from the problem of living and food, health issues have always been a concern for humanity. Various types of hereditary and epidemic diseases are the greatest adversaries of human health. The main cause of their generation and spread is different types of germs. Bacteria are one of the main causes of human health problems, and various medications have been used to treat it. Different treatments have been adopted to deal with the bacteria problems. The use of nanomaterial’s for therapeutic purposes, as in other areas of life, have been blessings and encouraging in recent times and is a hope for future. The current research work was focused to synthesize silver nanoparticles from lemon leaves extract and explore their antibacterial activity on different types of bacteria. Leaves extract worked as a reducing agent to form Ag^oatoms, which further clustered to form silver nanoparticles. These silver nanoparticles were characterized by using X-Ray Diffraction (XRD) analysis, UV-Visible Spectroscopy and Scanning Electron Microscopy (SEM). The UV-Visible absorption spectra of two different prepared silver nanoparticles samples were obtained and it was observed that these samples have prominent absorption response at energetic end of visible light from 380 nm to 410 nm. The crystalline nature of synthesized silver nanoparticles was confirmed by X-Ray diffraction analysis and showed the diffraction peaks at 31.86^o, 37.74^o, 43.3^o, 64.14^o, and 77.08^o, corresponding to silver element. The antibacterial activity was tested against different pathogenic bacteria and it was observed that the average inhibition zone diameter was 10 – 13 mm which shows the positive control. The green synthesis method is very economical, eco-friendly, and simple to implement for synthesis of silver nanoparticles

A HIGH PERFORMANCE HEXAGONAL MICROSTRIP PATCH ANTENNA WITH ENHANCED BANDWIDH AND VSWR FOR KU-BAND SATELLITE APPLICATIONS

Sat, 11 Apr 2026 00:00:00 +0500

One major reason for the increasing use of microstrip patch antennas in the existing satellite communication system is due to their compactness, lightweight nature, and ease of fabrication. In this study, an innovative microstrip patch antenna for satellite communication in Ku band was designed and analyzed. The design and simulation of the antenna were performed using CST studio suite software. The resonant frequency obtained from the antenna is 16.056 GHz, where the return loss and voltage standing wave ratio are –42.98 dB and 1.014, respectively. Furthermore, the antenna has bandwidth of 2.52 GHz, gain of 5.78 dBi, and directivity of 6.25 dBi. Based on comparative study of the designed antenna with the reference antenna, it can be concluded that the proposed antenna performs better than the reference antenna in terms of bandwidth and S-parameters

REVIEW ON DEEPFAKE TECHNOLOGY AND ITS IMPACT ON SOCIAL LIFE

Waqas Ahmad, Zeeshan, Syed M. Hassan Shah, Muhammad Irfan Fareed, Uzair Iqbal — Sat, 11 Apr 2026 00:00:00 +0500

Deepfake technology, which uses advanced artificial intelligence architectures like Generative Adversarial Networks (GANs), diffusion models, and transformer-based systems to produce highly realistic manipulated audiovisual content, is a major breakthrough in the creation of synthetic media. This study explores the technological development of deepfakes from lab testing to widely available apps, evaluating both their beneficial uses in the fields of entertainment, accessibility, education as well as its potential for detrimental abuse. This study shows how deepfake proliferation fundamentally challenges conventional assumptions regarding audiovisual evidence reliability and public trust in digital information ecosystems through an analysis of documented cases from 2025 involving corporate fraud exceeding $25 million, political disinformation campaigns, and widespread privacy violations targeting vulnerable populations.The study assesses current detection techniques, such as frequency-domain analysis, multimodal verification systems, and convolutional neural networks, while critically analyzing their limitations in generalizing beyond training datasets in the face of quickly developing generative capabilities. Additionally, this study examines ethical standards and regulatory frameworks that are developing in many jurisdictions, highlighting the need for coordinated worldwide governance structures that strike a balance between the advancement of technology and the defense of individual rights. The results show that integrated strategies combining technological breakthroughs in detection, extensive media literacy programs, cooperative governance standards, and strong legal protections are necessary for effective mitigation. In the end,In order to maintain information integrity in a time of synthetic media, this paper concludes that deepfake technology exhibits dual-use characteristics, where its societal impact is largely dependent on governance mechanisms, ethical deployment practices, and public awareness rather than inherent technological attributes. This calls for ongoing interdisciplinary collaboration.

INTRUSION DETECTION BASED ON FEDERATED LEARNING – A REVIEW

Abu Ubaida, Khurram Zeeshan Haider, Temur-ul-Hassan, Muhammad Azam Rasheed — Sat, 11 Apr 2026 00:00:00 +0500

The lack of access to raw data during model training is addressed by privacy-preserving methods, but the construction of intrusion detection systems remains essential for federated learning. The recent increase in FL-based intrusion detection research between 2019 and 2025 is reviewed, covering study metadata, data types, FL architectural designs, communication architectures, model design, and performance results. We include 90 central studies in five tables, showcasing their areas of application, data and preprocessing, FL topology, client diversity and aggregation strategies, model design with privacy in mind, and performance and robustness outcomes. Our findings show that IoT and IIoT applications, horizontal FL, and FedAvg are common; non-IID data and class imbalance are frequent; and public benchmarks like NSL-KDD and CICIDS2017 are widely used. Nevertheless, standardized FL-IDS benchmarks, energy and latency reporting as well as strong aggregation techniques are not well studied. We single out such promising directions as hierarchical and personalized FL, federated data augmentation, privacy- and robustness-oriented aggregation, and cross-dataset benchmarks. This review aims to assist the researchers and practitioners to swiftly realize the present state of the field, the most important gaps, and concentrate on research that can hasten the implementation of dependable FL-based intrusion detection.

SOLID STATE TRANSFORMERS AND WIRELESS POWER TRANSFER ARCHITECTURES FOR EXTREME FAST CHARGING OF ELECTRIC VEHICLES

Huzaifa Himad, Moiz Jillani, Umair Bhatti — Mon, 13 Apr 2026 00:00:00 +0500

The rapid global shift toward electric vehicles (EVs) has intensified the demand for extreme fast charging (XFC) systems capable of reducing charging duration without compromising grid stability, power quality, or system reliability. Two emerging technologies at the forefront of this transformation are Solid-State Transformers (SSTs) and Wireless Power Transfer (WPT), in both static and dynamic configurations. This work presents a technical analysis of state-of-the-art SST architectures and WPT compensation topologies designed for high-power EV charging applications. Key engineering trade-offs including efficiency, power density, cost, electromagnetic interference (EMI), thermal management, and reliability are examined in detail based on recent high-impact research.

Building on these insights, we propose a hybrid SST–WPT architecture that integrates a medium-voltage SST front-end with adaptive wireless charging pads/tracks, enabling both static and low-speed dynamic charging. A hypothetical 400-kW, 800-V XFC station is simulated using a cascaded H-bridge (CHB) rectifier and Dual Active Bridge (DAB) isolation stage coupled with a magnetically coupled resonant WPT system for pad-based charging. Simulation results indicate ~95–96% efficiency under ideal alignment and ~92% under moderate coil misalignment, fast DC-bus regulation, and controlled EMI behavior under realistic misalignment and parasitic conditions. The study also outlines open research challenges and future directions in coil alignment, MFT optimization, wide-bandgap reliability, EMI shielding, and multi-MW station scalability. The proposed framework demonstrates the technical viability of combining CHB, DAB, and resonant WPT technologies to enable modular, efficient, and grid-aware XFC infrastructure.

WATER HEATED HUMIDIFICATION-DEHUMIDIFICATION DESALINATION ACCOMPANYING AIR HEATING VIA EXIT BRINE

Tue, 14 Apr 2026 00:00:00 +0500

The pure water shortage around the globe is worrying for human beings, therefore extraction of pure water from seawater is achieved through desalination technologies. Thermal desalination processes are many, but humidification-dehumidification (HDH) desalination is preferable because of its simplicity and usable for a small scale unit, and easily maintainable. The HDH is based on evaporation and condensation of water vapors in humidifier and dehumidifier, respectively with carrying gas like air. The energy is needed for two purposes: water vapor evaporation (the latent heat of evaporation) while another purpose is to increase the ability of air to carry away those vapors in the humidifier. The water-heated and air-heated system have higher water productivity than the bare water or air heated system alone. Therefore, in this study, modeling and thermal analysis of a water-heated HDH system along with air pre-heating (before humidifier) was achieved by utilizing the brine heat energy. The model is based on thermodynamics principles, mass, and energy conservation laws. It was found that 3.6% of energy was recovered from brine and was utilized in heating the air before humidifier, which increased the inlet air temperature and hence enhanced the evaporation, and finally, condensation was increased and that was observed with 2.6% increase in the system productivity.

DESIGNING MOFS-BASED CATALYSTS FOR EFFICIENT CO2 CONVERSATION

Irfan Haider, Mujahid Abbas, Adeeb Ur Rehman, Kaleem Ullah — Tue, 14 Apr 2026 00:00:00 +0500

The development of efficient catalysts for CO₂ conversion is essential for mitigating climate change and enabling a circular carbon economy. In this study, metal–organic frameworks (MOFs)‑based catalysts are rationally designed and evaluated for efficient CO₂ conversion through electrocatalytic, photocatalytic, and thermocatalytic pathways. MOFs with high surface areas, open metal sites, and functionalized linkers are synthesized and characterized using X‑ray diffraction, nitrogen physisorption, and electron microscopy, confirming their crystalline nature, porosity, and structural integrity. CO₂ adsorption isotherms reveal strong host–guest interactions, with high uptake capacities and favorable binding energies that facilitate CO₂ activation. Electrocatalytic CO₂ reduction tests show enhanced activity and selectivity toward CO and formate, with Faradaic efficiencies exceeding 70–85% on optimized MOF‑based electrodes, while Cu‑rich MOFs and MOF‑derived catalysts promote C–C coupling to C₂+ products such as ethylene and ethanol. Photocatalytic and thermocatalytic experiments further demonstrate that MOF–semiconductor composites and amino‑functionalized MOFs can achieve high CO₂ conversion and product selectivity under mild conditions. The results establish clear structure activity relationships highlighting the role of tailored metal nodes, functional linkers, and embedded single‑ or dual‑atom sites in enhancing CO₂ conversion performance. These findings provide a framework for designing MOFs‑based catalysts that combine high efficiency, selectivity, and stability for practical CO₂ utilization technologies.

GREEN SYNTHESIS OF SILVER NANOPARTICLES USING PLANT EXTRACTS FOR ENHANCED ANTIMICROBIAL AND WOUND HEALING APPLICATIONS

Javeria Mehmood, Aijaz Ahmed Bhutto, Zain-ul-Abideen — Tue, 14 Apr 2026 00:00:00 +0500

Silver nanoparticles (AgNPs) have emerged as promising nanomaterials in biomedical applications due to their strong antimicrobial and wound healing properties. This study focuses on the eco-friendly green synthesis of AgNPs using plant extracts as natural reducing and stabilizing agents, addressing the limitations of conventional chemical synthesis methods that involve toxic reagents and environmental hazards. The methodology involved preparation of plant extract followed by its reaction with silver nitrate under controlled conditions to synthesize AgNPs. The nanoparticles were purified and characterized using advanced analytical techniques including UV–Visible spectroscopy, FTIR, XRD, SEM, and EDX to determine their optical, structural, and morphological properties. The results confirmed successful synthesis of crystalline, nanoscale AgNPs with a characteristic surface plasmon resonance peak at 404 nm. Biological evaluations demonstrated significant antimicrobial activity against Staphylococcus aureus, Escherichia coli, and Candida albicans, with a concentration-dependent increase in inhibition zones and low minimum inhibitory concentration (MIC) values. The nanoparticles also exhibited considerable antioxidant activity, achieving up to 78.6% radical scavenging efficiency. Cytotoxicity analysis revealed good biocompatibility, maintaining over 80% cell viability at concentrations ≤50 µg/mL. Furthermore, in vitro scratch assay results showed enhanced fibroblast migration and accelerated wound closure, reaching approximately 87.9% healing within 24 hours.

In conclusion, plant-mediated AgNPs demonstrate multifunctional properties including antimicrobial, antioxidant, and wound healing capabilities. The study highlights their potential as sustainable and effective alternatives for biomedical applications, particularly in wound management, while emphasizing the need for further in vivo investigations for clinical translation.

AUTOMATED SCALP DISORDER DIAGNOSIS USING ATTENTION-ENHANCED DEEP LEARNING WITH CLINICAL INTERPRETABILITY AND USABILITY VALIDATION

Hasnain Abdullah, Muhammad Awais, Riaz Ahmed — Tue, 14 Apr 2026 00:00:00 +0500

Hair and scalp disorders affect millions of individuals worldwide, resulting in significant physical, psychological, and social consequences. Traditional diagnostic methods rely on subjective visual assessment by dermatologists, which is time-consuming, error-prone, and often inaccessible in underserved regions. This paper presents a machine learning framework for the automated diagnosis of scalp disorders using non-invasive dermoscopic and trichoscopic imaging. The proposed system employs a convolutional neural network (CNN) architecture based on fine-tuned ResNet-50, augmented with U-Net segmentation, Gaussian denoising, and CLAHE contrast enhancement. A dataset of 5,000 labeled scalp images was used for training and evaluation, with a 70/15/15 train-validation-test split. The system achieved an overall accuracy of 92.5%, precision of 91.3%, recall of 93.2%, F1-score of 92.2%, and AUC of 0.96 — surpassing prior published approaches. A user study with 10 dermatologists confirmed the system matched expert diagnostic accuracy in 95% of cases. This work demonstrates the transformative potential of deep learning-based dermatological tools for equitable, scalable, and accurate healthcare delivery.

PERFORMANCE OPTIMIZATION OF 132 KV ELECTRICAL POWER GRID STATION

*M. A. Raza, Muhammad Shahid, Darakhshan Ara, Fatima Tul Zuhra — Tue, 14 Apr 2026 00:00:00 +0500

A case study of 132 KV grid station is under taken for improving the reliability, stability, and also for efficient operation of the power system. Electrical Transient Analyzer Program (ETAP) software is used to model the study which covers the three major aspects of power system such as Load Flow Analysis (LFA), Short Circuit Analysis (SCA), and Transient Stability Analysis (TSA). LFA identified under voltage and transformer overloading occurrences in the system. As a solution to these problems, a Static VAR Compensator (SVC) system was then incorporated in order to enhance voltage stability, decrease transformer loading and active and reactive power losses. SCA is used to assess the fault tolerance of the system by simulating different fault scenarios such as Line-to-Grounded (LG), Line-to-Line (LL), Double Line-to-Ground (LLG) and Three-phase (3Φ) faults. The results provided valuable information that enabled adequate understanding of fault currents and protective device coordination for safe and reliable system operation. The dynamic response of the system to disturbances, such as faults and recovery scenarios, will be assessed through TSA. The key parameters of the analysis were observed over a time interval of 60 seconds, which proved system stability under fault recovery, and provided important data for stability improvement in the future. This study not only overcame the current operational obstacles but also led to the optimization of the performance of substations.

Keywords : Power System Stability, Grid Station, Load Flow, Short Circuit, Transient Stability.

EXPLORING THE APPLICATIONS OF FRACTIONAL CALCULUS IN MODELING REAL-WORLD PHENOMENA

Faizan Ahmad, Sana Ramzan, Iqra Azeem — Wed, 15 Apr 2026 00:00:00 +0500

The paper looks into the application of the high order calculus in one of the real-world phenomena and its efficiency compared to the classical integer-order calculus. The results indicate that the fractional models make the prediction process more accurate by 21 percent making it between 68 and 89 percent and errors are less by a factor of over 50 percent, 32 to 14 percent. It is also demonstrated in the experiment that with the application of fractional calculus, it becomes possible to stabilize the system to a greater extent of 18 percent and the accuracy of long-term predictions is also enhanced by 62-84 percent. Fractional models have a very high memory effect representation (85%) when compared to classical models (45%), and indicate that they are able to model the complex dynamics of a system.

The breadth of the field of fractional calculus is illustrated by domain analysis showing 30, 27, and 25 percent performance improvements in biological systems, physics and engineering respectively. However, the study records an increase in computation time by 35 percent as among the limitations. Physics (62 percent) and engineering (58 percent) have the highest adoption rates.

Overall, the data helps to confirm the hypothesis that the notion of fractional calculus is more specific, coherent, and comprehensive in the modeling of complex systems. The study observes the need of improved computing methods and broader applications in order to achieve the full potential of the application of fractional-order models in science and engineering.

ENHANCING BUILDING ENERGY EFFICIENCY: INNOVATIVE CONCRETE WITH PHASE CHANGE MATERIAL-COATED COARSE AGGREGATE

Wed, 15 Apr 2026 00:00:00 +0500

Global warming presents a pressing global challenge, with the construction sector emerging as a significant contributor to carbon emissions, accounting for 25% to 40% of global carbon emissions in Pakistan. To address this issue, the construction industry is seeking innovative methods to enhance thermal insulation in buildings. Recent attention has been drawn to phase change materials (PCM) for their thermal storage properties, owing to challenges in electrification, rising costs of digital equipment, electric transportation, and residential heating and cooling. PCM, substances that absorb and release energy at phase transition temperatures, offer promising solutions for heating and cooling applications in construction materials. This study focuses on developing PCM-coated coarse aggregates, employing lauric acid and paraffin liquid through vacuum impregnation. Different ratios of PCM-coated aggregates replaced natural coarse aggregates in concrete mixes (0%, 25%, 50%, 75%, 100%). The study aimed to assess the mechanical, thermal, and bonding characteristics of concrete by evaluating slump value, compressive strength, and thermal conductivity. The results indicated that 50% PCM-coated aggregates have exhibited a consistent physical strength and excellent heat conduction across various applications of concrete. However, PCM-100% and PCM-75% compositions demonstrate heightened thermal energy conduction sensitivity but lower compressive strengths that make PCMs inappropriate for high-stress atmospheres. This research underscores the potential of PCM-embedded concrete for sustainable construction practices in the construction industry.

EXPLORING THE APPLICATIONS OF FRACTIONAL CALCULUS IN MODELING REAL-WORLD PHENOMENA

Faizan Ahmad, Sana Ramzan, Iqra Azeem — Wed, 15 Apr 2026 00:00:00 +0500

WATER HEATED HUMIDIFICATION-DEHUMIDIFICATION DESALINATION ACCOMPANYING AIR HEATING VIA EXIT BRINE

Wed, 15 Apr 2026 00:00:00 +0500

Laplacian Harmonic Spectral Analysis for IoT Network Topology Characterization

Mehtab Khan, Muhammad Uzair Khan, Sabir Shah, Asad Jan — Wed, 15 Apr 2026 00:00:00 +0500

The rapid expansion of the Internet of Things (IoT) has led to the deployment of large-scale, heteroge neous, and dynamically evolving network topologies. Understanding the structural properties of such net works is critical for ensuring robustness, efficient communication, and fault tolerance. Spectral graph theory provides powerful tools for analyzing network structures through eigenvalues of matrix representa tions. In this paper, we introduce the Laplacian Har monic matrix as an alternative spectral framework for modeling IoT network topologies. We investigate its fundamental properties and analyze the associ ated eigenvalues for various graph families. Further more, we interpret these spectral characteristics in the context of IoT networks, including star, mesh, and bipartite topologies. The results demonstrate that Laplacian Harmonic eigenvalues provide deeper insights into connectivity, robustness, and structural balance compared to traditional Laplacian matrices. The proposed framework offers a novel approach for IoT network design, monitoring, and optimization.

COTTON WEED DETECTION USING FASTER R-CNN ON COTTONWEEDDET3 DATASET FOR PRECISION AGRICULTURE

Thu, 16 Apr 2026 00:00:00 +0500

Weed species can significantly impact crop productivity, and manual weeding is often impractical due to labor intensity and scale. Consequently, many recent studies have focused on automating weed detection using image-based approaches. However, accurately detecting weeds through images remains a challenging task because the texture, color, and shape of weeds and crops are often very similar. In this study, we propose a deep learning-based solution using the Faster Region- Based Convolutional Neural Network (Faster R-CNN) architecture to detect three cotton weed species: carpetweed, morning-glory, and Palmer amaranth. We utilize a publicly available dataset, CottonWeedDet3, which contains 848 RGB images annotated with bounding boxes following the Common Objects in Context (COCO) format. Our proposed model achieved a mean Average Precision (mAP@0.5) of 92.3% at an Intersection Over Union (IoU) threshold of 0.5. The findings demonstrate the effectiveness of Faster R-CNN for accurate and auto- mated cotton weed detection in the context of precision agriculture.

INTELLIGENT ADAPTIVE MACHINE LEARNING SCALABLE FRAMEWORK FOR DYNAMIC MALWARE IDENTIFICATION AND PROACTIVE THREAT PREVENTION

Saleh Rehman, Sonia Jamil, Nasir Hussain, Assad Latif, Sohail Ahmad — Thu, 16 Apr 2026 00:00:00 +0500

Malware and computer viruses continue to pose serious threats to modern digital systems, often compromising security and leading to significant financial losses. As technology evolves, traditional malware detection methods—particularly signature-based approaches—are becoming less effective. These methods rely heavily on manual updates, respond only after threats are identified, and struggle to keep up with the growing volume and complexity of cyberattacks. In this study, we explore the potential of machine learning as a more advanced and proactive solution for malware detection. Unlike conventional techniques, machine learning models can learn from historical data and identify patterns that indicate malicious behavior. This capability allows them to detect previously unknown or emerging malware variants without requiring explicit signatures. The research focuses on evaluating the performance of three widely used machine learning algorithms: Random Forest, Gradient Boosting, and Support Vector Machine. These models are analyzed based on key feature sets and performance metrics to determine their effectiveness in identifying malicious software. Experimental findings demonstrate that these algorithms significantly improve detection accuracy while reducing false positives.Furthermore, the adaptability of machine learning models enables continuous improvement as new data becomes available, making them highly suitable for dynamic threat environments. The proposed approach also supports real-time detection, which is critical for minimizing damage caused by fast-spreading malware. In addition, the scalability of these techniques allows them to be implemented across large and complex networks without major performance degradation. This study also highlights the importance of feature selection and data preprocessing in improving model efficiency and accuracy. By integrating intelligent detection mechanisms, organizations can strengthen their cybersecurity infrastructure and respond more effectively to evolving threats. Overall, this research contributes to the development of robust, flexible, and sustainable malware detection systems, offering a promising direction for future advancements in cybersecurity.

PREDICTIVE ANALYTICS FOR CUSTOMER CHURN IN SUBSCRIPTION-BASED BUSINESSES USING MACHINE LEARNING

Farwa Zainab, Farwa Nazim, Muhammad Kashaf, Naeem Aslam, Muhammad Sajid Maqbool — Thu, 16 Apr 2026 00:00:00 +0500

This research paper focuses on developing a robust framework for predicting customer churn in subscription-based industries. Using the uploaded thesis data, we implemented various machine learning algorithms, including Logistic Regression, Random Forest, and Gradient Boosting. The study emphasizes the importance of data preprocessing and feature engineering. Our findings indicate that ensemble methods provide the highest predictive performance with an accuracy of 88.7% and a superior ROC-AUC score of 0.95. The analysis further highlights that 'Customer Age,' 'Active Membership Status,' and 'Number of Products' are the most significant predictors of churn. This proposed system provides a scalable and adaptive early-warning framework, enabling businesses to implement proactive, data-driven retention strategies to maximize customer lifetime value.

TOWARDS IDENTIFICATION AND REFINEMENT BY USING PATTERN-BASED APPROACH

Rizwana Yasmeen, Dr. Mudasir Mahmood — Thu, 16 Apr 2026 00:00:00 +0500

In contemporary (Modern computing systems support the execution of complex business processes through advanced software engineering methods. One of the key challenges in this context is ensuring that requirements are properly identified and refined to reflect the trust expectations of end-users throughout the software development life cycle. Trust has become an essential quality attribute for contemporary applications, making it necessary to integrate user concerns into early design stages. This research introduces a structured set of user-friendly questions aimed at helping end-users clearly express their trust-related needs. In addition, a pattern-based strategy is applied to the requirements engineering phase, offering practitioners a systematic approach to capture, evaluate, and address trust concerns as part of the development process.

Keywords : Identification, requirements, refinement, patterns, MSDA (modern software development approaches)

HYBRID LIGHTNING PROTECTION DESIGN FOR WIND–SOLAR RENEWABLE ENERGY SYSTEMS UNDER EXTREME WEATHER CONDITIONS

Hafiz Muhammad Azib Khan, Waqas Arif — Tue, 14 Apr 2026 00:00:00 +0500

Background: Hybrid wind-solar renewable energy systems (HRES) are becoming more popular as a means of generating renewable energy in a sustainable way. Nonetheless, lightning strike is susceptible to these systems particularly during extreme weather conditions and it presents a danger to equipment and operational dependability. Direct and induced lightning strike damage should be minimized by proper lightning protection.

Purpose: The purpose of the study is to create a hybrid type of lightning protection in wind-solar energy systems to increase the resiliency of the device to extreme weather conditions and lightning-related damages.

Procedure: The study employs a risk-based analysis, which involves simulations and field experiments to develop and test a lightning protection system. Mathematical modeling of the electrical and electromagnetic behavior of the system is done using simulation tools such as MATLAB/Simulink and COMSOL Multiphysics. Field tests are done to test the design performance against the reality that includes transient voltage suppression and surge protection.

Findings: The findings demonstrate that combined measures of lightning protection, such as surge protecting devices (SPDs), air terminals, and optimized grounding systems can significantly decrease downtimes of the systems and maintenance expenses. Field tests proved that the control system reduced downtime by 90% and the maintenance cost of the system by 35%.

Conclusion: The proposed hybrid design of lightning protection is effective in mitigating lightning hazards, which guarantees wind-solar systems long-term and safe operation during extreme weather conditions.

BANDWIDTH CONSUMPTION AND LATENCY TRADE OFF IN VIDEO STREAMING OVER SDN (QOE BASED ANALYSIS)

Fri, 17 Apr 2026 00:00:00 +0500

This paper examines the trade-off between bandwidth usage and latency in video streaming using Software Defined Networking (SDN), which is a limitation of the current literature since it considers the values of these two parameters separately instead of being joint. The research will seek to investigate their joint influence on Quality of Experience (QoE) and give a more comprehensive perspective on streaming performance. The SDN-based controlled SDN-based testbed was used as a quantitative experimental design with the implementation in Mini-net and the Ryu controller. VLC was used in video streaming, and network conditions like bandwidth (1-3 Mbps), latency and network packet loss were varied systematically. The metrics of QoE such as the startup delay and buffering incidents were measured and studied. The findings indicate that high-resolution video (38402160) took up to approximately 3 seconds to start and buffering at 1 Mbps, whereas low-resolution video (640360) took approximately 1 second before starting up and buffering without buffering. At 3 Mbps, high-resolution streams did experience continuous buffering, which validated that bandwidth and latency affect QoE in concert. The empirical evidence of bandwidthlatency interaction in SDN settings presented in the study and the necessity of combined QoE optimization frameworks.