https://jemsc.qom.ac.ir/ Journal of Engineering Management and Soft Computing en daily 1 Wed, 01 Apr 2026 00:00:00 +0330 Wed, 01 Apr 2026 00:00:00 +0330 https://jemsc.qom.ac.ir/article_4312.html I am happy to present the second issue of Volume 12 of the Journal of Engineering Management and Soft Computing (JEMSC) for April 2026. This issue features a diverse collection of twelve original research articles that cover the journal’s main areas: engineering management, soft computing, optimization, and intelligent decision-making. The papers in this issue tackle complex challenges across various sectors, including manufacturing, telecommunications, finance, healthcare, and supply chain management. Several contributions focus on optimization frameworks. These include a multi-objective mathematical model for shop floor flow production using the Grey Wolf metaheuristic, a new Society Deciding Process algorithm for multi-objective problems, and hybrid approaches that combine mathematical modeling with machine learning for cloud manufacturing scheduling. In the realm of data-driven intelligence, we see innovative applications, such as a smart banking facility model that integrates big data and macroeconomic variables, a federated learning-based data governance framework for fraud detection, and a wavelet-enhanced GARCH-deep learning hybrid for predicting stock market volatility. Other important works investigate risk analysis in 5G radio access networks, investment security clustering using multi-period multi-criteria decision-making, reverse logistics network design for used goods, and a robust adaptive inverse control scheme for vehicle path tracking. In the biomedical field, a dual-branch hybrid-feature learning network provides interpretable support for pediatric colonoscopy decisions. We conclude this issue with a scenario-based robust decision-making model for health tourism service supply chains, reflecting the journal’s commitment to applied engineering management. I sincerely thank the authors for their valuable contributions, the reviewers for their careful evaluations, and the editorial team for their hard work. I believe the diverse and high-quality research presented here will spark new ideas and encourage more interdisciplinary collaboration. We invite researchers, academics, and postgraduate students to submit their work that is relevant and aligned with the journal’s focus. We look forward to seeing the interesting submissions that will help enhance JEMSC’s role in the scientific community. Sincerely, Jalal Rezaeenour, Editor-in-Chief Journal of Engineering Management and Soft Computing (JEMSC) https://jemsc.qom.ac.ir/article_4293.html The flow shop production system has attracted a lot of attention in research, and many researchers have conducted research to optimize flow shop production systems. The important point is to correctly assign activities to each machine in order to minimize the time to complete all activities. In the present study, a model for optimizing the number of human activities in a flow shop production system is presented, which is based on a two-objective model. The result of solving the model shows that the mathematical model has a very good ability to solve the problem in small dimensions. The gap between the results of the deterministic solution and the metaheuristic model has been reported to be zero percent. However, considering that in small and medium problems, the gap between the mathematical model and the metaheuristic model was negligible, the results presented are reliable by relying on the results calculated according to the metaheuristic model. https://jemsc.qom.ac.ir/article_4294.html The fifth generation of mobile networks (5G) offers new and advanced services such as virtual/augmented reality (AR/VR), high-quality video streaming, remote surgery, Internet of Things (IoT), and smart cars with stringent requirements. In this article, we review the 5G architecture and then focus on the Radio Access Network (RAN) architecture, and collect the threats, vulnerabilities, and security solutions provided by researchers and academic authorities in this field. In this article, we also rank RAN threats in the 5G network. To rank threats, we assign a score to each threat by introducing a risk criterion, and finally, we analyze the risk by assessing the impact/severity and probability of success of the threats on the 5G network, so that we can provide a good ranking for the threats. The results of our studies and analyses in this article inform 5G network operators which threats should be prioritized. The results of this paper will guide to secure the 5G network and deploy security solutions. https://jemsc.qom.ac.ir/article_4295.html Accurate volatility forecasting is vital for effective decision-making in financial markets, yet remains a complex challenge due to the noisy, non-stationary nature of financial time series and the diversity of volatility definitions. This study proposes a novel hybrid framework that combines statistical, signal processing, and machine learning techniques to enhance volatility prediction. The approach begins with wavelet transformations to extract multi-scale features from raw financial data, effectively addressing non-stationarity. These features are then evaluated using multiple volatility estimators to determine their predictive relevance. The framework integrates GARCH models, wavelet-derived inputs, and deep learning architectures, with Particle Swarm Optimization (PSO) employed for optimal parameter tuning. Leveraging S&P 500 data from 2000 to 2024 and incorporating multi-source inputs, the model achieves a more holistic representation of market dynamics. Empirical results demonstrate that the hybrid method significantly reduces prediction errors and consistently outperforms baseline models and established benchmarks. To validate its practical utility, we developed trading strategies based on the predicted volatility. Backtesting results show substantial performance gains. https://jemsc.qom.ac.ir/article_4296.html Advancements in technology and the emergence of multi-objective optimization problems across various scientific domains have spurred research and development of novel metaheuristic algorithms to address these challenges. Although these methods have largely succeeded in approaching the Pareto-optimal front, the optimization process has not been fully realized. This paper introduces a multi-objective version of the Social Division Process (SDP) algorithm, termed MOSDP, aimed at improving the quality of Pareto front solutions. The MOSDP algorithm employs a memory structure as an archive to store non-dominated solutions. Additionally, it utilizes a non-dominated sorting mechanism based on crowding distance to establish a hierarchical social division structure and guide the evolutionary process in the multi-objective problem space. The performance of MOSDP is evaluated using 18 well-known multi-objective test functions, UF, and IMOP, and compared with the Multi-Objective City Councils Evolution (MOCCE), Multi-Objective Ant Lion Optimization (MOALO), Multi-Objective Slime Mould Algorithm (MOSMA), and Multi-Objective Artificial Hummingbird Algorithm (MOAHA). The results of the Friedman average rank test demonstrate the superiority of MOSDP over the aforementioned algorithms in terms of Inverted Generational Distance (IGD), Generational Distance (GD), and Maximum Spread (MS) metrics. https://jemsc.qom.ac.ir/article_4297.html The granting of bank facilities and the targeting and identification of suitable customers by banks is a serious concern. The non-repayment of granted loans by customers can severely damage the profitability of banks, leading them to move towards proprietorship and economic activity, which ultimately results in increased inflation and many other economic problems. Based on the problem mentioned, the aim of the present research is to design an intelligent model for granting bank facilities based on big data, considering macroeconomic variables, sanctions, and economic shocks. To design this model, 6 macroeconomic variables, shocks, and sanctions were included in the model. The model was evaluated using four machine learning algorithms: multiple regression, support vector machine, decision tree, and random forest, based on customer data from the country's banks. Subsequently, economic growth, unemployment rate, and Gini coefficient have relatively less influence, estimated to affect loan repayment or the number of unpaid loans by 9%, 10%, and 11% respectively. https://jemsc.qom.ac.ir/article_4298.html With the growing volume of data interactions and increasing complexity of oversight in executive agencies, the need for innovative approaches to fraud detection and data governance has become more pressing than ever. Given that data is stored in a distributed manner, cloud computing emerges as an effective solution. However, security concerns hinder direct data exchange between organizations. To address this challenge, the present study proposes a decentralized, data-driven governance model based on federated learning and cloud infrastructure. In this model, each organization preprocesses its data at the edge and extracts fraud-related features. These results are then transmitted to a central server, where deep learning techniques are used to predict new inter-organizational fraud patterns. This approach preserves confidentiality and enables collaborative analysis without requiring data aggregation. Experimental results show that the proposed method reduces computational complexity by 60% and achieves a fraud detection accuracy of 97.6%, demonstrating its high effectiveness in multi-organizational and distributed environments. https://jemsc.qom.ac.ir/article_4299.html Investment security is a critical component of economic development in countries, as its enhancement fosters investor confidence and encourages participation in productive sectors and financial markets. Given the regional disparities across the country, assessing investment security at the provincial level is essential. This study aims to rank and cluster Iranian provinces in terms of investment security over the period 2021–2023, using a multi-criteria, multi-period decision-making approach. For this purpose, two techniques—MP-TOPSIS and MULTI-MOORA—were employed to evaluate and rank the provinces, and the results were compared. Subsequently, the k-means clustering method was applied to group provinces into homogeneous clusters. The data were obtained from 12 policy reports published by the Iranian Parliament Research Center, and the weighting of criteria was performed using three methods: MEREC, Shannon entropy, and CRITIC. The findings reveal that Semnan, Golestan, and Hormozgan provinces exhibit the highest levels of investment security, while Tehran, Sistan and Baluchestan, and Kohgiluyeh and Boyer-Ahmad rank lowest. Additionally, the correlation between the two ranking methods was estimated at 87%, indicating a high degree of consistency and validating the proposed model. The results suggest that macroeconomic stability, administrative transparency, and robust legal frameworks are the most influential factors in determining investment security. Region-specific policy recommendations based on these analyses can significantly improve the investment climate across the country https://jemsc.qom.ac.ir/article_4300.html This research investigates multi-objective, multi-level scheduling within a cloud manufacturing environment, employing metaheuristic algorithms. Initially, a comprehensive literature review was conducted, followed by the development of a multi-objective mathematical model integrated with a machine learning (ML) model. The model’s validity was first assessed by solving it for small-scale instances. Given that the exact method was only feasible up to the tenth instance, metaheuristic algorithms were utilized for solving the model in larger dimensions. The results demonstrated the model’s solvability in large-scale scenarios using the NSGAII algorithm. Subsequently, the model was solved considering risk input values, revealing that among the temporal parameters, transportation time and setup time exhibit the most significant impact on overall time. Among the cost-related parameters, preparation cost has the greatest effect on time, with transportation cost being the next most significant. Crucially, activity time emerges as the most impactful cost parameter, with transportation cost following as the second most influential. https://jemsc.qom.ac.ir/article_4301.html In supply chains, the goal is usually to meet demand at the lowest cost. But there are cases where either the transportation costs are insignificant or, such as in critical situations, the supply of more goods has a much higher priority than the costs. In such cases, instead of minimizing the cost, we should maximize the transfer flow values. In this case, the supply chain network minimization problem (minimum cost flow) becomes a type of flow maximization problem (maximum flow). In this paper, we have addressed a type of flow maximization problem in supply chains. First, we have defined and modeled it, then, considering its complex structure, we have obtained a suitable approximate solution for it by using a meta-heuristic method. https://jemsc.qom.ac.ir/article_4302.html Colorectal neoplasia, while rare, is high-risk in children. As such, colonoscopic detection of paediatric polyps is of clinical importance. Current deep learning–based computer-aided systems are trained on adult data and do not provide solutions for class imbalance, cross-centre generalisation or interpretability in paediatrics. We introduce an imbalance-aware dual-branch hybrid-features (IDHF) framework for polyp versus non-polyp classification. IDHF, based on a ResNet-50 backbone, is equipped with complementary texture branch to augment deep semantic features, adaptive gating for late fusion and optimisation of a class-weighted loss function with false-negative penalty and branch-agreement regularisation. The proposed model is trained on CP-CHILD-A (8,000 images, 1,000 polyp/7,000 non-polyp) and tested without fine-tuning on CP-CHILD-B (1,500 images, 400 polyp/1,100 non-polyp) on another platform. ResNet-50+IDHF achieves 99.6% accuracy, 100.0% sensitivity, 99.5% specificity, 96.8% precision and a 98.4% F1-score on CP-CHILD-A. ResNet-50+IDHF achieves 99.5% accuracy, 99.2% sensitivity, 99.6% specificity, 99.0% precision and a 99.1% F1-score on CP-CHILD-B, supporting a robust and interpretable solution for computer-aided paediatric polyp detection. https://jemsc.qom.ac.ir/article_4303.html Compared to the other research that focuses on the intelligent identification of nonlinear systems, the deliberated methodology evolved the inverse intelligent process as a universal controller for a class of nonlinear systems and this method has achieved a remarkably low root mean square (RMS) and tracking error rate. In the vehicle systems, the tracking of the predefined desired path is challenging to achieve, especially in presence of disturbances. The planned AIDS procedure consists of an online inverse model identifier updated using the back-propagation (BP) algorithm. In this approach, an offline identification phase provides the initial network weights. A Multilayer Perceptron (MLP) is then employed as a nonlinear controller, trained to represent the system's inverse dynamics and applied to the vehicle model. The convergence of the noisy states to the nominal ones, the robustness of the recommended designing system, and the reduction of noisy phenomena effect in the system's state are all crucial advantages of the planned AIDC. Simulation results demonstrate the promising performance of the proposed methodology. https://jemsc.qom.ac.ir/article_4304.html Health tourism is a dynamic sector of Iran’s economy with strong potential in the religious city of Qom. This study develops plausible future scenarios for the health tourism service supply chain and identifies suitable strategies using Robustness Analysis (RA). A mixed qualitative–quantitative design is employed. First, key influencing factors are identified through expert interviews and a literature review. Next, alternative scenarios are constructed and evaluated using a weighted decision-matrix that simultaneously accounts for system complexity and environmental uncertainty in a participatory decision-making setting. The analysis yields seven distinct scenarios derived from eight key factors, ranging from an ideal future characterized by sustainable development and attraction of religious markets to a pessimistic future marked by sanctions-driven stagnation and multiple crises. The RA results indicate that vertical integration and market development perform best under favorable conditions, whereas retrenchment and heterogeneous diversification are more effective under crisis conditions. The proposed weighted RA decision matrix provides a practical decision-support tool for policymakers and healthcare managers in Qom to select resilient strategies across diverse environmental futures. https://jemsc.qom.ac.ir/article_3860.html Although deep neural networks (DNNs) achieve remarkable accuracy, their deployment on energy-constrained edge devices remains a significant challenge due to their intricate architectures and high computational complexity. To mitigate these limitations, this study introduces a novel energy-aware multi-exit Vision Transformer (ViT) architecture tailored for energy-harvesting edge environments. The proposed model incorporates an early-exit mechanism, enabling dynamic termination of inference at intermediate layers depending on the current energy availability. A weighted loss function is developed to facilitate joint optimization of both intermediate and final outputs, wherein exponential weighting is employed to enhance the performance of earlier exits. Experimental evaluations reveal that the proposed method substantially reduces computational overhead while maintaining competitive final accuracy. Specifically, utilizing the fourth exit results in a 28.5% reduction in FLOPs with minimal degradation in accuracy. These results underscore the potential of the proposed architecture to achieve an effective trade-off among accuracy, latency, and energy efficiency in dynamic inference scenarios on resource-limited edge platforms. https://jemsc.qom.ac.ir/article_3869.html The purpose of this research is to design a financial technology valuation model in the solar energy industry. For this purpose, the opinions of the experts of the regional electricity distribution company of Isfahan province were collected and a qualitative model based on the data theory of the foundation was developed. The model includes three main categories (technological valuation, technology organization and investment) and 9 sub-categories (low cost, exclusive experience, value provided by technology, communication, technological infrastructure, agility, knowledge management, human resources and new financing model). Content validity analysis has shown that all categories are sufficient to be included in the valuation model. Using the structural equation model in Lisrel software, the relationships of indicators and the influence of each factor have been investigated. The results show that the relationship between human resources and technological organization is meaningless and therefore it can be removed. Finally, distribution functions have been evaluated with Decision Tools software and the influence of various factors in 90, 80 and 70 percent scenarios. https://jemsc.qom.ac.ir/article_3872.html This research simultaneously investigates the problems of multi-project scheduling, supplier selection, material ordering, and warehouse location and capacity selection. Given the cost trade-offs among these issues, solving them separately leads to suboptimal solutions. For each type of material, multiple suppliers with quantity-dependent discount policies exist. Several potential locations for establishing project material storage warehouses have been identified. A mathematical model is presented to solve these problems simultaneously, with the objective function minimizing the total cost. Since this problem is NP-hard, a genetic algorithm is proposed, and its parameters are tuned using the Taguchi method. Finally, a sample example is solved in both integrated and segregated cases, and the results are compared to demonstrate the model's efficiency. Sensitivity analysis showed that the variability in project activity durations, the rewards for early completion and penalties for project delays, the suppliers' discount policies, and the nonlinearity of transportation costs have the most significant impact on the savings resulting from solving the integrated model. https://jemsc.qom.ac.ir/article_3965.html In modern software systems characterized by high complexity and dynamic behavior, behavior-based regression testing plays a crucial role in ensuring reliability following system modifications. This paper presents a novel event-driven framework for behavior-oriented regression testing that leverages reinforcement learning to optimize test selection. The proposed architecture models the system's behavior as sequences of events extracted from logs and interactions, enabling a learning agent to derive effective policies for identifying behavioral anomalies and regressions based on environmental feedback. The framework is implemented within a simulated environment, and experimental results demonstrate its superiority over traditional approaches by reducing test costs and improving the precision of critical regression detection. This study introduces an innovative approach to automated software testing and highlights the role of reinforcement learning in enhancing the quality and stability of complex systems.