As the issue of agricultural sustainability has continued to increase, there has been a need to use data based solutions to improve agricultural productivity. This paper proposes a machine learning system combining Random Forest and XGBoost to combine prediction-forecasting crop yield and classification of crop type rice and wheat in Indian state of Uttarakhand. The model is tested using a library of 6, 000 samples containing 12 soil and climatic characteristics and measured on regression and classification quality. The hybrid ensemble with hyperparameter optimization and verified on the basis of 10-fold cross-validation performed better than single base learners in all measures. It achieved a classification accuracy of 96.3 and R2 = 0.927. Statistically significant developments that were formed using paired t-tests were set at p = 0.05. The SHAP and ablation analysis found out nitrogen, rainfall, and pH as the most influential features. The forecasted framework provides a better generalizability, interpretability, and computational effectiveness, which is appropriate to be applied in the designs of real life in precision agronomy. The new result is novel, interpretable, and high-performative to crop yield intelligence in data-scarce areas and provides a contribution to this study.
The study compares the capabilities of various time series and machine learning models including ARIMA, LSTM, CatBoost, XGBoost, and LightGBM, by predicting the equity movement for major Indian infrastructure and energy companies with hydrogen related exposure, namely Larsen & Toubro, NTPC Limited, JSW Energy Limited, and Adani Green Energy Limited. Hydrogen fuel is considered the most promising energy provider of the future, and an understanding of its position in the market is vital for its growth. The study uses historical data involving stock prices from April 2019 through April 2024 obtained from the National Stock Exchange of India. Using open price as the primary variable, the performance of the models is measured. Additional variables such as close price, highest price, lowest price, and volume are used for gradient boosting. Output graphs comparing actual prices and predicted prices are obtained. The results indicate that deep learning and gradient boosting outperform the statistical model. LSTM demonstrated the strongest short-term predictive accuracy through sequential learning among all models. Among the gradient boosting models, LightGBM provides consistent and robust performance by effectively capturing nonlinear feature interactions. Overall, the study highlights the growing importance of machine learning in interpreting India’s renewable energy equity markets.
Reliable and efficient functioning of thermal power plants is necessary for a steady power supply and economic viability. In the present research, the reliability and availability of a subsystem of a thermal power plant are modelled through a Continuous-Time Markov Process (CTMP), which reflects the stochastic change from working to failed states. Human error is incorporated as an additional failure factor to reflect practical operating conditions. To further improve performance, a Particle Swarm Optimization (PSO) algorithm, a nature-inspired metaheuristic approach, is used to maximize system availability. In addition to enhancing availability, the research uses PSO to maximize the expected profit of the system using a unified economic objective function. The Markov-based model yields an initial system availability of 0.9117. After applying Particle Swarm Optimization, the availability improves to 0.9199 with variation in population size and further increases to 0.9240 with variation in the number of iterations, representing an overall enhancement of approximately 1.23%. The optimized results also increase the expected profit. The outcomes demonstrate that the integration of Markov modelling with PSO ensures accurate reliability and availability analysis and provides a robust framework for economic optimization.
Medical imaging is essential for obtaining precise information to assess patient health and provide effective treatment. The initial analysis of a wide range of medical images is vital for identifying abnormalities. However, limitations during the image acquisition process can lead to poor image quality. To address the issues of reduced information and contrast in medical images, the proposed method is designed. The method integrates the spatial frequencies (SF) of both the original image and the image enhanced through Triple Clipped Dynamic Histogram Equalization (TCDHE). Additionally, the discrete wavelet transform and singular value decomposition are applied simultaneously to both images to obtain an improvement factor (gamma), which controls the contrast enhancement rate. The use of spatial frequencies helps preserve the detailed components of an image, such as edges and sharp features. For experimental purposes, three types of datasets are utilized, including MRI, X-ray, and ultrasound. Objective evaluation is performed using seven performance metrics: AMBE, PSNR, SSIM, GMSD, REC, entropy, and subjective evaluation with a mean opinion score. The experimental results demonstrate that AMBE (4.08), SSIM (0.99), PSNR (37.67 dB), and GMSD (0.13) are the best among state-of-the-art methods. However, the results for REC and entropy are comparable to those of state-of-the-art methods. Furthermore, the average values of all performance parameters have been computed across three categories of medical datasets to demonstrate the efficacy of the proposed method.
Supply Chains (SC) are undergoing rapid transformation with the advent of Industry 4.0 (I4.0) technologies. In highly dynamic and volatile market conditions, Supply Chain Flexibility (SCF) that is the capability to adapt and respond to dynamic changes in the market environment and Supply Chain Sustainability (SCS) that is the ability of managing SCs to meet sustainability requirements and improve performance have become strategically critical. Despite growing interest, limited research has quantitatively examined which factors create the greatest synergy between SCF and SCS. This study addresses this gap by quantitatively prioritizing Organizational Practices (OP) and Underlying Drivers (UD) in the SC of pharmaceutical context using the Analytic Hierarchy Process (AHP). Findings reveal that OP such as ‘Integration of sustainability principles into SCF’ and ‘Instrumental, interconnected and intelligent SC’ holds the highest relative importance, demonstrating its dominant influence on the decision context and UD such as ‘Transparency, Traceability and Visibility’ holds the highest relative significance, exhibiting its dominant effect in the decision environment, followed by ‘Technological changes/ adoption’ and ‘Resilience’ have the highest impact on achieving simultaneous improvements in flexibility and sustainability. The results provide a structured decision hierarchy, highlighting actionable areas for managerial investment and resource allocation. By identifying and ranking these key enablers, the study demonstrates that SCF and SCS can co-evolve rather than compete, offering both practical guidance for pharmaceutical managers and a foundation for future research on optimizing SC performance in the digital era.
Massive multiple-input multiple-output (mMIMO) systems are the backbone of modern-day wireless communication due to their potential to utilize spectrum efficiently and increase network capacity. Secure and optimal beam selection is key to interference and security challenges in dense urban areas, especially with the arrival of 5G and beyond. Classical solutions such as exhaustive beam search or even statistical models have high computational complexity and are not so flexible to dynamic situations. This study generated simulated data based on the realistic distribution of the users and the phenomena of the terrestrial. The data set records common metrics like the user locations, channel states, and beamforming settings. We propose a deep-learning framework that predicts top-K transmit–receive beam pairs using only receiver location and then enforces physical-layer security (PLS) by filtering out pairs that violate an eavesdropper-power threshold. On simulated DeepMIMO-inspired scenes, our model attains Top-1/Top-5/Top-10 accuracies of 69.51%/85.32%/92.43%, cuts beam-search overhead by 92.19%, and reduces mean execution time to 95 ms. With security constraints, it approaches achievable bounds for Probability of Successful Detection (PSD)/ Probability of Secure Signal Detection (PSSD) and reduces estimated eavesdropping probability from 15.6% to 5.2%, while improving secrecy capacity and Bit Error Rate (BER). The novelty is a security-constrained beam selection loop integrated directly into initial access, requiring low CSI and remaining deployable within 5G NR procedures.
This paper presents a novel multi-criteria decision-making (MCDM) framework based on interval-valued Fermatean fuzzy sets (IVFFSs) to effectively handle ambiguity and imprecision inherent in real-world decision environments. When accurate assessments are not accessible, the suggested model offers a more adaptable and realistic representation of ambiguous information by expanding classical Fermatean fuzzy theory through interval-valued membership and non-membership degrees. In order to aggregate expert evaluations, a new group generalized interval-valued Fermatean fuzzy weighted average (GGIVFFWA) operator is developed. This operator allows for the simultaneous examination of many advisers and deciders perspectives and incorporates group-based parameters. The applicability and efficiency of the proposed framework are demonstrated through a real-world case study on strategic partner selection for credit risk assessment. The obtained outcomes confirm the robustness, stability, and reliability of the proposed technique, which are further validated through a comparative analysis with existing decision-making approaches. General, the proposed framework offers an efficient and practical tool for solving complex MCDM problems under uncertainty.
In computational imaging, multi-focus image fusion is a critical process that aims to produce a single image that covers all-in-focus areas from numerous partially focused input images. In this paper, we present a novel approach using a Super-Resolution Generative Adversarial Network (SRGAN) specifically designed for multi-focus image fusion. First, we create a new multi-focus image dataset from the publicly accessible COCO dataset. This process generates a complete collection of annotated image pairs with different focus areas. The generator is designed using a residual learning architecture and upsampling layers. The generator creates a high-resolution fused image with features and texture preservation by processing two input images. Using PatchGAN-based implementation, the discriminator ensures that the fused images maintain global and local coherence through adversarial training. Putting emphasis on intensity, structural similarity, and perceptual qualities, we combine content loss with adversarial loss to achieve balanced learning. Extensive trials on public multi-focus image datasets show that our SRGAN-based model achieves superior fusion quality and texture consistency by outperforming five current state-of-the-art approaches in both quantitative and visual evaluations. The proposed method achieves real-time performance, meets the requirements of contemporary image fusion applications, and demonstrates its efficacy in generating high-quality fused images.
Sustainable transportation has led to a surge in zero-carbon engine technologies. The environmental impact of internal combustion (IC) engines has become a crucial concern. To meet Euro 7 emission norms and fuel economy, the Exhaust Gas Recirculation (EGR) system technology has emerged as a promising technology for enhancing engine efficiency and reducing nitrogen oxide (NOx) emissions. This paper covers the following essential aspects. The Euro 7 emission norms are more stringent than Euro 6, by 35-56 percent for NOx and 13-27 percent for particulate matter (PM). EGR rate can vary between 10 percent and 60 percent, depending on the emission-reduction strategy adopted by the engine manufacturer. Almost 70 percent of diesel vehicles adopted EGR technology over other technologies. The market trend is shifting toward hybrid EGR-plus-SCR and EGR-plus-LNT technologies. Cooled EGR technology appears to be a standard adaptation for gasoline engines rather than an optional technology for fuel economy, with 2-10 percent fuel economy gains depending on the engine manufacturer's adaptation. Based on comprehensive technological advances over the past two decades in diesel engines, the study proposed that cooled EGR is a promising technology for NOx reduction up to 80 percent, hinting at the future of automotive technology to meet Euro7 emission norms.
Federated learning enables distributed devices to train a shared model without transmitting raw data to a central server. In real-world networks, devices intermittently disconnect and reconnect in bursts. When a device returns after a prolonged offline period, its update is computed from an outdated global model; such stale updates introduce noise, increase gradient variance, and slow convergence. Existing client selection methods either ignore staleness or address it post hoc through aggregation, and therefore fail to jointly optimize staleness and bursty connectivity. This paper proposes SAB-Select, a client selection method that scores each client using three signals: staleness, burst availability, and gradient diversity. Theoretical convergence analysis demonstrates that SAB-Select reduces the number of rounds required to reach a target accuracy. This paper also introduces an optional audit log for accountability, which can be instantiated as a signed append-only log or as a permissioned blockchain ledger. Experiments on MNIST and Fashion-MNIST demonstrate that SAB-Select reaches 85% accuracy faster (7 vs. 12 rounds), reduces average staleness, and maintains fairness comparable to the baselines. A cost-based analysis (communication bytes and time proxy) further demonstrates that faster convergence translates into reduced bandwidth and latency requirements for reaching the target accuracy. The results demonstrate that staleness-aware client selection provides a practical, theoretically grounded solution for federated learning on realistic edge networks.
The integration of home solar photovoltaic (PV) systems with electric vehicle (EV) charging infrastructure has gained more popularity with the advancement towards low-carbon electric vehicles and localized solar energy systems. Although the use of electric vehicles and solar PV systems has been widely researched, the integration of these two has not been studied much, particularly in the context of emerging economies' densely populated urban centers. Using the mixed-method, user-centric approach, the study has initiated the analysis with the use of focus groups with important stakeholders to identify the key factors for the integration of solar PV systems with electric vehicle charging infrastructure. Causal interdependencies are then formulated, and these factors are prioritized under uncertainty with the use of the multi-criteria decision-making framework, namely, Grey DEMATEL. The results indicate the key driving factors such as improved customer comfort, advancements in the use of renewable energy, and the availability of government incentives. Conversely, the barriers to the integration of solar PV systems with electric vehicle charging infrastructure are found to be the limited energy storage capacity, the lack of home charging infrastructure, and the lack of solar energy availability during off-peak hours. This study has highlighted the policy interventions to improve the solar PV charging barriers with the use of streamlined subsidies, effective public education, and advancements in technology, which could be useful for the wider acceptance of solar-powered electric vehicle charging infrastructure, thereby achieving the sustainability objectives.
The human factor plays a decisive role in the safe and reliable functioning of modern smart grids, where increasing automation and system complexity impose greater cognitive load on operators. This paper presents a hybrid approach that combines the Fuzzy Analytic Hierarchy Process (FAHP) and Bayesian Network (BN) to assess the probability of human error under uncertain conditions. Triangular fuzzy numbers are used to express expert judgments and are processed through FAHP to obtain the relative significance of key performance shaping factors (PSFs). The normalized weight of each subfactor is calculated and converted into fuzzy possibility scores (FPSs), which are further transformed into fuzzy failure probabilities (FFPs). These probabilities are incorporated into a BN model built in GeNIe to evaluate the effects of individual and grouped factors on human reliability. The model indicates a high level of operator reliability while highlighting the need for improvement in critical situations. The findings suggest that training and knowledge sharing programs, alarm design and management, cognitive load, and shift management are among the most influential subfactors. A case study from the smart grid power distribution sector was conducted to demonstrate the applicability of the proposed framework. Expert opinions from experienced power system engineers and grid operators were collected using purposive sampling to evaluate human reliability factors. The hybrid framework provides a clear understanding of how different conditions affect operator reliability and supports utilities in improving training, communication processes, and control room design. The proposed FAHP BN framework offers a systematic and flexible method for analyzing human reliability in the smart grid environment and serves as a practical tool for identifying critical factors and guiding actions that enhance the safety and reliability of smart grid operations.
This study's primary aim is to understand the role and interaction of various sustainability criteria within industrial symbiosis networks, and it embarks upon this task by way of the Decision-Making Trial and Evaluation Laboratory (DEMATEL) method. This study seeks to establish the main driving forces and dependent factors that could affect the success and resiliency of these networks, enhancing sustainable industrial practices. A strong decision-making tool, DEMATEL is used in this study to visualize complex cause-and-effect relationships between sustainability criteria through a systematic matrix approach. Data was collected from experts around industrial symbiosis, covering diverse sectors for comprehensive sustainability criteria. The analysis quantitatively assesses influence and dependency among identified criteria, clearly depicting systemic interactions. It illustrates that the criteria mentioned, resilience in IS networks, adoption of innovative technologies, and optimization of material exchanges are important driving forces in the system that can put considerable influence on other criteria. Conversely, the criteria of waste reduction and environmental impact reduction are highly dependent and are greatly changed by the actions of these drivers. This study thus emphasizes the crucial balance between the influential and dependent criteria, which offers insights for designing targeted strategies to strengthen the weaker areas and weaken the strong drivers to boost overall system sustainability. This manuscript presents a unique application of the DEMATEL methodology in the Turkish textile sector; an application not previously explored in the country’s industry. Hence, mapping the complex cause-and-effect relationships connecting sustainability criteria in industrial symbiosis answers a great concern in the existing literature.
It is worth mentioning that Blockchain Technology can enhance transparency, trust, and efficiency in a supply chain in general and in food supply chains in particular. Nonetheless, its application in Iran’s food industry is still limited due to technological, organizational, strategic and operational challenges. Unlike previous studies, this paper identifies and structures the key factors influencing blockchain implementation in the food sector supply chain through expert confirmation and Interpretive Structural Modeling (ISM). Twenty-six critical drivers were analyzed to reveal their interdependencies and hierarchical significance. The results of this work highlight that developing efficient and effective food supply chain strategies is the most critical factor for successful blockchain adoption, supported by components such as technological readiness, management commitment, transparency, trust, and system quality. The findings also provided a strategic roadmap, which emphasizes aligning technology adoption with organizational preparedness and policy support. This research contributes to the understanding of blockchain diffusion in emerging economies and offers practical and actionable insights for decision-makers seeking to enhance traceability, efficiency, and stakeholder trust in the complex operations of the food supply chain.
Quality control in the pharmaceutical industry has always faced a fundamental challenge in balancing product release speed, safety, operational costs, and regulatory requirements. In many pharmaceutical organizations, the final decision for batch release still relies on extensive human reviews and conservative procedures, which leads to operational bottlenecks and reduced organizational agility. Despite recent advances in digital twin and risk-based approaches, a unified and operational framework for realizing quality control without direct human intervention has not been systematically developed. This research presents, for the first time, a closed-loop framework for touchless quality control (Touchless QC) based on the integration of hybrid digital twin, fuzzy multi-objective optimization model, and human intervention mechanism under exceptional conditions. In this framework, the digital twin of the laboratory flow simulates the propagation of quality risk and resource constraints, and the multi-objective optimization engine simultaneously minimizes the release time, cumulative quality risk, quality control costs, human intervention rate, and laboratory congestion under GMP and SLA constraints. The meta-heuristic algorithms NSGA-II, MOEA/D, and MOPSO are used to extract the Pareto front. Results from real and simulated industrial data show that the proposed framework is able to significantly reduce the release time and quality control costs, while maintaining the quality risk level at an acceptable level or even lower than traditional policies. Also, release decisions are made in a structured and automated manner, and human intervention is activated only when risk or uncertainty thresholds exceeds the defined limits. In addition to improving operational performance, this framework also enhances transparency, traceability, and regulatory robustness of decisions, and is a practical step towards realizing smart, risk-based quality control in the pharmaceutical industry.
This paper presents an approximation method of integrable functions using a modified Barbosu operator, aimed at improving the rate of convergence in function approximation on the interval [0,1]. By introducing a suitable adjustment in the weight function, we construct a sequence of positive linear operators that better preserve the function's characteristics and demonstrate superior approximation behaviour, are studied. Theoretical error bounds are established in terms of the first and second order modulus of smoothness using their equivalence with the K-functionals. Finally, numerical experiments using test functions validate the theoretical findings and confirm that the King-type modified Barbosu operator achieve better approximation performance than the usual Kantorovich type Barbosu operator.
The healthcare sector relies on inventory planning to ensure the continuous availability of emergency medications while minimizing wastage and expiry. This study focuses on enhancing the total profitability by jointly optimizing order quantity and replenishment cycle time of the pharmaceutical products. A differential equation is developed to represent time-dependent demand and product expiration, and the resulting cost function of the inventory is transformed into a non-linear minimization problem. Convexity of the model is examined through graphical pairwise analysis of the decision variables. The model considers practical industry factors of the pharmaceutical sector, including the time-sensitive demand, shelf life and trade credit. The optimality of the proposed model is evaluated using conventional and metaheuristic algorithms such as CA, CSA, and RSA. Numerical experiments based on real data on credit period are collected from local and chain pharmacies, demonstrating the applicability of the proposed model. This study offers a framework for pharmaceutical inventory management, providing actionable insights to enhance operational efficiency and profitability while managing time-sensitive healthcare products.
Excessive advancement in technology and the Internet of Things (IoT) have brought a revolution in Cloud Computing. However, a massive amount of data is generated from IoT devices, ultimately affecting the Cloud’s efficiency. Fog Computing has been developed to improve efficiency. It places Fog nodes near the IOT devices to reduce the processing latency. Fog computing has achieved remarkable success; however, its nodes are resource-constrained, which makes efficient Task-scheduling. The proper scheduling of tasks among appropriate Fog nodes, considering the energy efficiency and reliability in terms of failure rate, is the main challenge for researchers. Many task scheduling algorithms have been proposed in the literature for energy efficiency and makespan; however, very little attention is paid to reliability. The high computational time of the task scheduling algorithm is another crucial aspect of these algorithms. This paper proposes a Reliable Hybrid Pareto-based Multi-objective (RHPMO) Task Scheduling approach based on metaheuristics. It combines two advanced metaheuristic techniques, named JAYA and Genetic Algorithm, to perform optimal task scheduling based on the Pareto front and explore the best global solution. A multi-objective function is formulated to minimize makespan, energy, and failure rate, thereby increasing the efficiency and reliability of task scheduling. Extensive experimentation is conducted on MATLAB R2023a on an Intel i3, 8 GB RAM. The simulation experiment results of the proposed hybrid approach are compared with the various metaheuristic approaches like JAYA, Genetic Algorithm, Particle Swarm Optimization, Bees Algorithm, and Ant Colony Optimization, and hybrid GA and PSO. The proposed approach surpasses various state-of-the-art studies and demonstrates an impressive improvement of 68.91% in computational time, 42.48% in makespan, 29.83% in energy consumption, and 12.58% in reliability, respectively.
The Newell-Whitehead-Segel (NWS) equation plays a significant role in nonlinear systems, including mathematical biology, plasma physics, solid-state physics, optics, quantum mechanics, cosmology, fluid dynamics, and many others. In this work, we proposed the quintic B-spline collocation method to find the numerical solution of the nonlinear NWS-type equation. Crank-Nicolson finite difference method (FDM) is used to discretize the equation in time space, and quasi-linearization is employed to linearize the nonlinear term. The stability analysis has been discussed using the Von Neumann Method, and stability conditions have been obtained. The numerical results are compared with existing techniques, which demonstrate the effectiveness and applicability of the proposed technique. The proposed method has been applied to four numerical test problems at various time levels and mesh sizes to demonstrate the effectiveness, which involves quadratic, cubic, and quartic order nonlinear terms. The comparison shows good agreement with the exact solution, as demonstrated by absolute error tables and graphs. Moreover, the proposed method is easy to implement and produces good results.
This paper solves the two-parameter singularly perturbed Fredholm integro-differential equations through the developed exponentially fitted operator method and monotone finite difference method. The differential component is determined computationally using a developed exponentially fitted operator approach and a monotone finite difference method. The composite trapezoidal rule evaluates the integral component on a uniform grid. The developed exponentially fitted operator method gives first-order convergence when the small parameter related to the perturbation is much smaller than the square of the second parameter, and second-order convergence when the square of the second parameter is much smaller than the perturbation parameter. In addition, the monotone finite difference method shows second-order convergence in both situations. Numerical results are included to support the theoretical findings of the proposed methods.
