Pub Date : 2025-07-28DOI: 10.1016/j.jocs.2025.102683
Xiaohua Zhang , Yujie Fan
Magnetohydrodynamics (MHD) has extensive applications in diverse fields, making the study of three-dimensional (3D) MHD problems crucial. For MHD flows, when the Hartmann () number is large, leading to a convection-dominated regime where convection terms overcome diffusion. In such scenarios, standard Galerkin methods fail to suppress non-physical oscillations in solutions, as they lack inherent stabilization mechanisms for strong convection. This paper introduces the variational multiscale element free Galerkin (VMEFG) method to solve 3D steady MHD equations. The VMEFG method inherits the advantage of the element free Galerkin (EFG) method in avoiding the complex meshing process, which is particularly challenging for complex 3D problems. Moreover, compared with the EFG method, it shows enhanced stability in dealing with convection-dominant problems and can automatically generate stabilized parameters, outperforming other stabilization techniques. To verify the numerical stability and accuracy of the VMEFG method, several numerical experiments on various domains including cubic, annular cubic, spherical, and annular spherical domains were conducted and compared with EFG solutions and existing literature results. The results indicate that the VMEFG method can effectively avoid numerical oscillations and maintain stability for 3D MHD problems at high number, providing a reliable and efficient solution for such problems.
{"title":"The variational multiscale element free Galerkin method for three-dimensional steady magnetohydrodynamics duct flows","authors":"Xiaohua Zhang , Yujie Fan","doi":"10.1016/j.jocs.2025.102683","DOIUrl":"10.1016/j.jocs.2025.102683","url":null,"abstract":"<div><div>Magnetohydrodynamics (MHD) has extensive applications in diverse fields, making the study of three-dimensional (3D) MHD problems crucial. For MHD flows, when the Hartmann (<span><math><mrow><mi>H</mi><mi>a</mi></mrow></math></span>) number is large, leading to a convection-dominated regime where convection terms overcome diffusion. In such scenarios, standard Galerkin methods fail to suppress non-physical oscillations in solutions, as they lack inherent stabilization mechanisms for strong convection. This paper introduces the variational multiscale element free Galerkin (VMEFG) method to solve 3D steady MHD equations. The VMEFG method inherits the advantage of the element free Galerkin (EFG) method in avoiding the complex meshing process, which is particularly challenging for complex 3D problems. Moreover, compared with the EFG method, it shows enhanced stability in dealing with convection-dominant problems and can automatically generate stabilized parameters, outperforming other stabilization techniques. To verify the numerical stability and accuracy of the VMEFG method, several numerical experiments on various domains including cubic, annular cubic, spherical, and annular spherical domains were conducted and compared with EFG solutions and existing literature results. The results indicate that the VMEFG method can effectively avoid numerical oscillations and maintain stability for 3D MHD problems at high <span><math><mrow><mi>H</mi><mi>a</mi></mrow></math></span> number, providing a reliable and efficient solution for such problems.</div></div>","PeriodicalId":48907,"journal":{"name":"Journal of Computational Science","volume":"91 ","pages":"Article 102683"},"PeriodicalIF":3.1,"publicationDate":"2025-07-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144713416","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-07-26DOI: 10.1016/j.jocs.2025.102678
Ning Ning Chung , Hamed Taghavian , Mikael Johansson , Lock Yue Chew
We study how neural networks can perform the task of elevator dispatching of commuters from their origins to their destinations. Instead of applying a neural network in the conventional way, we construct a specific neural network architecture that optimizes the commuters’ traveling time after taking into account the domain knowledge and the efficacy of potential future actions. The constructed architecture is modular with building blocks of neuronal structure that serve specified functional roles. By relaxing the weights and then training this network via reinforcement learning, we show that it outperforms an agent that implements the standard elevator algorithm. More remarkably, we observe the spontaneous emergence of functional modules within the structure of the network in consequence of the action sequences experienced during training. This behavioral feature of the neural network makes it less of a black box, with specific aspects of its functions being explicitly discernible from its network connections.
{"title":"A demonstration on the construction of modular neural network using elevator system that operates based on reinforcement learning","authors":"Ning Ning Chung , Hamed Taghavian , Mikael Johansson , Lock Yue Chew","doi":"10.1016/j.jocs.2025.102678","DOIUrl":"10.1016/j.jocs.2025.102678","url":null,"abstract":"<div><div>We study how neural networks can perform the task of elevator dispatching of commuters from their origins to their destinations. Instead of applying a neural network in the conventional way, we construct a specific neural network architecture that optimizes the commuters’ traveling time after taking into account the domain knowledge and the efficacy of potential future actions. The constructed architecture is modular with building blocks of neuronal structure that serve specified functional roles. By relaxing the weights and then training this network via reinforcement learning, we show that it outperforms an agent that implements the standard elevator algorithm. More remarkably, we observe the spontaneous emergence of functional modules within the structure of the network in consequence of the action sequences experienced during training. This behavioral feature of the neural network makes it less of a black box, with specific aspects of its functions being explicitly discernible from its network connections.</div></div>","PeriodicalId":48907,"journal":{"name":"Journal of Computational Science","volume":"91 ","pages":"Article 102678"},"PeriodicalIF":3.7,"publicationDate":"2025-07-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144724832","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-07-26DOI: 10.1016/j.jocs.2025.102682
Nitu Kumari, Anurag Singh
A significant challenge in various fields of science and engineering is extracting governing equations from data. Prey-predator models are particularly complex due to their nonlinear behavior, making traditional analytical methods insufficient for accurately capturing their dynamics. In this study, we introduce a data-driven approach to model the intricate dynamics of Hastings–Powell model solely from time series data. This article explores the application of the sparse identification of nonlinear dynamics (SINDy) and its extension, the SINDy-PI (parallel, implicit) method, in a model representing a chaotic food chain. The main goal is to determine the governing equations that describe the chaotic dynamics of the prey-predator populations. Hence, this study uses the parameters wherein the dynamics exhibit chaotic behavior. The method of SINDy was developed with the aim of identifying governing equations of nonlinear dynamical systems. In both methods, a library of potential terms are created and then a regression problem is solved. We have employed both methods as our model incorporates not only nonlinear terms but also rational terms. Our results shows that SINDy method is unable to find the exact form of governing equations but SINDy-PI method has the capability to accurately capture the authentic structure of the governing equations. In addition, we applied model selection techniques to identify the most parsimonious model possible. Through the application of SINDy and SINDy-PI, this research contributes to the advancement of data-centric approaches in ecological modeling, offering insights into the intricate dynamics of multi-species interactions within ecosystems. Further, for this study to be more realistic, utilizing real-world data from three-species would have been ideal. However, due to non-availability of three species real data, simulated data set has been used for validation purpose.
{"title":"Data-driven enhancement of the Hastings–Powell model using sparse identification algorithm","authors":"Nitu Kumari, Anurag Singh","doi":"10.1016/j.jocs.2025.102682","DOIUrl":"10.1016/j.jocs.2025.102682","url":null,"abstract":"<div><div>A significant challenge in various fields of science and engineering is extracting governing equations from data. Prey-predator models are particularly complex due to their nonlinear behavior, making traditional analytical methods insufficient for accurately capturing their dynamics. In this study, we introduce a data-driven approach to model the intricate dynamics of Hastings–Powell model solely from time series data. This article explores the application of the sparse identification of nonlinear dynamics (SINDy) and its extension, the SINDy-PI (parallel, implicit) method, in a model representing a chaotic food chain. The main goal is to determine the governing equations that describe the chaotic dynamics of the prey-predator populations. Hence, this study uses the parameters wherein the dynamics exhibit chaotic behavior. The method of SINDy was developed with the aim of identifying governing equations of nonlinear dynamical systems. In both methods, a library of potential terms are created and then a regression problem is solved. We have employed both methods as our model incorporates not only nonlinear terms but also rational terms. Our results shows that SINDy method is unable to find the exact form of governing equations but SINDy-PI method has the capability to accurately capture the authentic structure of the governing equations. In addition, we applied model selection techniques to identify the most parsimonious model possible. Through the application of SINDy and SINDy-PI, this research contributes to the advancement of data-centric approaches in ecological modeling, offering insights into the intricate dynamics of multi-species interactions within ecosystems. Further, for this study to be more realistic, utilizing real-world data from three-species would have been ideal. However, due to non-availability of three species real data, simulated data set has been used for validation purpose.</div></div>","PeriodicalId":48907,"journal":{"name":"Journal of Computational Science","volume":"91 ","pages":"Article 102682"},"PeriodicalIF":3.7,"publicationDate":"2025-07-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144724831","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-07-25DOI: 10.1016/j.jocs.2025.102684
Jan Trynda , Paweł Maczuga , Albert Oliver-Serra , Luis Emilio García-Castillo , Robert Schaefer , Maciej Woźniak
Despite their flexibility and success in solving partial differential equations, Physics-Informed Neural Networks (PINNs) often suffer from convergence issues, even failing to converge, particularly in problems with steep gradients or localized features. Several remedies have been suggested to solve this problem, but one of the most promising is the dynamical adaptation of the collocation points. This paper explores a novel adaptive sampling method, of a stochastic nature, based on the Adaptive Mesh Refinement used in the Finite Element Method. The error estimates in our refinement algorithm are based on the value of the residual loss function. We tested our method against a variety of 1D and 2D benchmark problems that exhibit steep gradients near certain boundaries, with promising results.
{"title":"An h-adaptive collocation method for Physics-Informed Neural Networks","authors":"Jan Trynda , Paweł Maczuga , Albert Oliver-Serra , Luis Emilio García-Castillo , Robert Schaefer , Maciej Woźniak","doi":"10.1016/j.jocs.2025.102684","DOIUrl":"10.1016/j.jocs.2025.102684","url":null,"abstract":"<div><div>Despite their flexibility and success in solving partial differential equations, Physics-Informed Neural Networks (PINNs) often suffer from convergence issues, even failing to converge, particularly in problems with steep gradients or localized features. Several remedies have been suggested to solve this problem, but one of the most promising is the dynamical adaptation of the collocation points. This paper explores a novel adaptive sampling method, of a stochastic nature, based on the Adaptive Mesh Refinement used in the Finite Element Method. The error estimates in our refinement algorithm are based on the value of the residual loss function. We tested our method against a variety of 1D and 2D benchmark problems that exhibit steep gradients near certain boundaries, with promising results.</div></div>","PeriodicalId":48907,"journal":{"name":"Journal of Computational Science","volume":"91 ","pages":"Article 102684"},"PeriodicalIF":3.7,"publicationDate":"2025-07-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144722460","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-07-25DOI: 10.1016/j.jocs.2025.102681
Juraj Kardoš , Wouter Edeling , Diana Suleimenova , Derek Groen , Olaf Schenk
Sensitivity analysis is an important tool used in many domains of computational science to either gain insight into the mathematical model and interaction of its parameters or study the uncertainty propagation through the input–output interactions. In many applications, the inputs are stochastically dependent, which violates one of the essential assumptions in the state-of-the-art sensitivity analysis methods. Consequently, the results obtained ignoring the correlations provide values which do not reflect the true contributions of the input parameters. This study proposes an approach to address the parameter correlations using a polynomial chaos expansion method and Rosenblatt and Cholesky transformations to reflect the parameter dependencies. Treatment of the correlated variables is discussed in context of variance and derivative-based sensitivity analysis. We demonstrate that the sensitivity of the correlated parameters can not only differ in magnitude, but even the sign of the derivative-based index can be inverted, thus significantly altering the model behavior compared to the prediction of the analysis disregarding the correlations. Numerous experiments are conducted using workflow automation tools within the VECMA toolkit.
{"title":"Sensitivity analysis of high-dimensional models with correlated inputs","authors":"Juraj Kardoš , Wouter Edeling , Diana Suleimenova , Derek Groen , Olaf Schenk","doi":"10.1016/j.jocs.2025.102681","DOIUrl":"10.1016/j.jocs.2025.102681","url":null,"abstract":"<div><div>Sensitivity analysis is an important tool used in many domains of computational science to either gain insight into the mathematical model and interaction of its parameters or study the uncertainty propagation through the input–output interactions. In many applications, the inputs are stochastically dependent, which violates one of the essential assumptions in the state-of-the-art sensitivity analysis methods. Consequently, the results obtained ignoring the correlations provide values which do not reflect the true contributions of the input parameters. This study proposes an approach to address the parameter correlations using a polynomial chaos expansion method and Rosenblatt and Cholesky transformations to reflect the parameter dependencies. Treatment of the correlated variables is discussed in context of variance and derivative-based sensitivity analysis. We demonstrate that the sensitivity of the correlated parameters can not only differ in magnitude, but even the sign of the derivative-based index can be inverted, thus significantly altering the model behavior compared to the prediction of the analysis disregarding the correlations. Numerous experiments are conducted using workflow automation tools within the VECMA toolkit.</div></div>","PeriodicalId":48907,"journal":{"name":"Journal of Computational Science","volume":"91 ","pages":"Article 102681"},"PeriodicalIF":3.7,"publicationDate":"2025-07-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144739413","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-07-24DOI: 10.1016/j.jocs.2025.102675
Yongsung Kwon , Minjin Lee , Mi Jin Lee , Seung-Woo Son
Understanding the dynamics of traffic clusters is crucial for enhancing urban transportation systems, particularly in managing congestion and free-flow states. This study applies computational percolation theory to analyze the formation and growth of traffic clusters within urban road networks, using high-resolution taxi data from Chengdu, China. Presenting the road network as a time-dependent, weighted, directed graph, we identify distinct behaviors in traffic jam and free-flow clusters through the growth patterns of giant connected components (GCCs). A persistent gap between GCC size curves, especially during rush hours, highlights disparities driven by spatial traffic correlations. These are quantified through long-range weight-weight correlations, offering a novel computational metric for traffic dynamics. Our approach demonstrates the influence of network topology and temporal variations on cluster formation, providing a robust framework for modeling complex traffic systems. The findings have practical implications for traffic management, including dynamic signal optimization, infrastructure prioritization, and strategies to mitigate congestion. By integrating graph theory, percolation analysis, and traffic modeling, this study advances computational methods in urban traffic analysis and offers a foundation for optimizing large-scale transportation systems.
{"title":"A computational analysis of traffic cluster dynamics using a percolation-based approach in urban road networks","authors":"Yongsung Kwon , Minjin Lee , Mi Jin Lee , Seung-Woo Son","doi":"10.1016/j.jocs.2025.102675","DOIUrl":"10.1016/j.jocs.2025.102675","url":null,"abstract":"<div><div>Understanding the dynamics of traffic clusters is crucial for enhancing urban transportation systems, particularly in managing congestion and free-flow states. This study applies computational percolation theory to analyze the formation and growth of traffic clusters within urban road networks, using high-resolution taxi data from Chengdu, China. Presenting the road network as a time-dependent, weighted, directed graph, we identify distinct behaviors in traffic jam and free-flow clusters through the growth patterns of giant connected components (GCCs). A persistent gap between GCC size curves, especially during rush hours, highlights disparities driven by spatial traffic correlations. These are quantified through long-range weight-weight correlations, offering a novel computational metric for traffic dynamics. Our approach demonstrates the influence of network topology and temporal variations on cluster formation, providing a robust framework for modeling complex traffic systems. The findings have practical implications for traffic management, including dynamic signal optimization, infrastructure prioritization, and strategies to mitigate congestion. By integrating graph theory, percolation analysis, and traffic modeling, this study advances computational methods in urban traffic analysis and offers a foundation for optimizing large-scale transportation systems.</div></div>","PeriodicalId":48907,"journal":{"name":"Journal of Computational Science","volume":"91 ","pages":"Article 102675"},"PeriodicalIF":3.1,"publicationDate":"2025-07-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144711667","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-07-24DOI: 10.1016/j.jocs.2025.102679
Saiful Islam , Md. Nahid Hasan , Pitambar Khanra
The increasing prevalence of graph-structured data across various domains has intensified greater interest in graph classification tasks. While numerous sophisticated graph learning methods have emerged, their complexity often hinders practical implementation. In this article, we address this challenge by proposing a method that constructs feature vectors based on fundamental graph structural properties. We demonstrate that these features, despite their simplicity, are powerful enough to capture the intrinsic characteristics of graphs within the same class. We explore the efficacy of our approach using three distinct machine learning methods, highlighting how our feature-based classification leverages the inherent structural similarities of graphs within the same class to achieve accurate classification. A key advantage of our approach is its simplicity, which makes it accessible and adaptable to a broad range of applications, including social network analysis, bioinformatics, and cybersecurity. Furthermore, we conduct extensive experiments to validate the performance of our method, showing that it not only reveals a competitive performance but in some cases surpasses the accuracy of more complex, state-of-the-art techniques. Our findings suggest that a focus on fundamental graph features can provide a robust and efficient alternative for graph classification, offering significant potential for both research and practical applications.
{"title":"A structural feature-based approach for comprehensive graph classification","authors":"Saiful Islam , Md. Nahid Hasan , Pitambar Khanra","doi":"10.1016/j.jocs.2025.102679","DOIUrl":"10.1016/j.jocs.2025.102679","url":null,"abstract":"<div><div>The increasing prevalence of graph-structured data across various domains has intensified greater interest in graph classification tasks. While numerous sophisticated graph learning methods have emerged, their complexity often hinders practical implementation. In this article, we address this challenge by proposing a method that constructs feature vectors based on fundamental graph structural properties. We demonstrate that these features, despite their simplicity, are powerful enough to capture the intrinsic characteristics of graphs within the same class. We explore the efficacy of our approach using three distinct machine learning methods, highlighting how our feature-based classification leverages the inherent structural similarities of graphs within the same class to achieve accurate classification. A key advantage of our approach is its simplicity, which makes it accessible and adaptable to a broad range of applications, including social network analysis, bioinformatics, and cybersecurity. Furthermore, we conduct extensive experiments to validate the performance of our method, showing that it not only reveals a competitive performance but in some cases surpasses the accuracy of more complex, state-of-the-art techniques. Our findings suggest that a focus on fundamental graph features can provide a robust and efficient alternative for graph classification, offering significant potential for both research and practical applications.</div></div>","PeriodicalId":48907,"journal":{"name":"Journal of Computational Science","volume":"91 ","pages":"Article 102679"},"PeriodicalIF":3.7,"publicationDate":"2025-07-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144779729","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-07-23DOI: 10.1016/j.jocs.2025.102677
Abu Safyan Ali , Muhammad Awais , Shumaila Javeed
Smoking dynamics created a global health crisis with major socioeconomic repercussions. It presents one of the most pressing issues the world has encountered for decades, affecting the social fabric, economy, and health globally. Sufficient treatment plans paired with significant coverage on radio, in print media, and social media as information sources may cause people to become more aware of the risks caused by smoking due to which individuals change their behavior and attitude toward smoking dynamics. In this study, we propose novel deterministic models for analyzing and controlling smoking dynamics. The model classifies the total population into five distinct sub-populations. Initially, we implement treatment for smokers, then the impact of media coverage of smokers on a daily basis along with proper treatment of smokers applies, and last one is about the combined effectiveness of TV, Radio, and all social media platforms (SMP) advertisement and treatment to addicted smokers. The disease-free equilibrium (DFE) and endemic equilibrium (EEP) states of proposed model one are qualitatively formulated, with stability analyses indicating local stability of DFE when 0 and of EEP when 0 . Global stability of the steady states is further examined using the Lyapunov function and Castillo-Chavez theorems. Sensitivity analysis of models is evaluated through the Normalized Sensitivity Index and Partial Rank Correlation Coefficient (PRCC) techniques. Furthermore, numerical simulations are used to verify the theoretical predictions of the proposed deterministic models. The simulation results suggest that targeted media coverage across different sources, including conventional and social media, together with competent medical care by treatment, may successfully lower the incidence of smoking. Through the use of awareness campaigns and advertising slogans, we can greatly increase public knowledge and eventually encourage quitting smoking.
{"title":"Mathematical modeling of smoking addiction control: Impact of treatment, news, and media campaigns","authors":"Abu Safyan Ali , Muhammad Awais , Shumaila Javeed","doi":"10.1016/j.jocs.2025.102677","DOIUrl":"10.1016/j.jocs.2025.102677","url":null,"abstract":"<div><div>Smoking dynamics created a global health crisis with major socioeconomic repercussions. It presents one of the most pressing issues the world has encountered for decades, affecting the social fabric, economy, and health globally. Sufficient treatment plans paired with significant coverage on radio, in print media, and social media as information sources may cause people to become more aware of the risks caused by smoking due to which individuals change their behavior and attitude toward smoking dynamics. In this study, we propose novel deterministic models for analyzing and controlling smoking dynamics. The model classifies the total population into five distinct sub-populations. Initially, we implement treatment for smokers, then the impact of media coverage of smokers on a daily basis along with proper treatment of smokers applies, and last one is about the combined effectiveness of TV, Radio, and all social media platforms (SMP) advertisement and treatment to addicted smokers. The disease-free equilibrium (DFE) and endemic equilibrium (EEP) states of proposed model one are qualitatively formulated, with stability analyses indicating local stability of DFE when <span><math><mi>R</mi></math></span> <sub>0</sub> <span><math><mrow><mo><</mo><mn>1</mn></mrow></math></span> and of EEP when <span><math><mi>R</mi></math></span> <sub>0</sub> <span><math><mrow><mo>></mo><mn>1</mn></mrow></math></span>. Global stability of the steady states is further examined using the Lyapunov function and Castillo-Chavez theorems. Sensitivity analysis of models is evaluated through the Normalized Sensitivity Index and Partial Rank Correlation Coefficient (PRCC) techniques. Furthermore, numerical simulations are used to verify the theoretical predictions of the proposed deterministic models. The simulation results suggest that targeted media coverage across different sources, including conventional and social media, together with competent medical care by treatment, may successfully lower the incidence of smoking. Through the use of awareness campaigns and advertising slogans, we can greatly increase public knowledge and eventually encourage quitting smoking.</div></div>","PeriodicalId":48907,"journal":{"name":"Journal of Computational Science","volume":"91 ","pages":"Article 102677"},"PeriodicalIF":3.7,"publicationDate":"2025-07-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144722461","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-07-23DOI: 10.1016/j.jocs.2025.102688
Hale Hapoglu , Egemen Ander Balas , Semin Altuntaş
The employment of stirred tank reactors in the field of treatment technology is well-established. In this regard, a bioreactor model is commonly utilized for conducting simulations, identifying parameters, and developing control applications. Control of biomass concentration is independent of scale through manipulation of the dilution rate. To enable discrete-time control, an equivalent model incorporating a zero-order hold element and a 0.1-h sampling time has been formulated for controlling biomass concentration. In this study, the various well-known controllers performed effectively to track set points. Further, to mitigate the effects of load disturbances, the generalized predictive controller, the proportional integral derivative controller, and the controllers designed based on pole placement have been employed to obtain process control responses. The performance of these controllers has been evaluated through a weighted aggregate sum product assessment technique that employs an analytical hierarchy process. Due to the significant nonlinearity present in the closed loop bioprocess with substrate inhibition, the feedforward artificial neural network controller is trained using a closed-loop dataset, and its performances are compared with the conventional controllers. The controller has demonstrated its ability to manage realistic feed fluctuations without risking upset to the culture. The biomass concentration showed only minor deviations, settling swiftly back to the desired value by smoothly adjusting the dilution rate. This controller with tansig and purelin functions overcomes nonlinearities and time delays better than conventional controllers. The results suggest that the artificial neural network controller offers the desired simplicity and effectiveness for industrial applications.
{"title":"Artificial neural network discrete-time biomass controller for a continuous stirred tank reactor","authors":"Hale Hapoglu , Egemen Ander Balas , Semin Altuntaş","doi":"10.1016/j.jocs.2025.102688","DOIUrl":"10.1016/j.jocs.2025.102688","url":null,"abstract":"<div><div>The employment of stirred tank reactors in the field of treatment technology is well-established. In this regard, a bioreactor model is commonly utilized for conducting simulations, identifying parameters, and developing control applications. Control of biomass concentration is independent of scale through manipulation of the dilution rate. To enable discrete-time control, an equivalent model incorporating a zero-order hold element and a 0.1-h sampling time has been formulated for controlling biomass concentration. In this study, the various well-known controllers performed effectively to track set points. Further, to mitigate the effects of load disturbances, the generalized predictive controller, the proportional integral derivative controller, and the controllers designed based on pole placement have been employed to obtain process control responses. The performance of these controllers has been evaluated through a weighted aggregate sum product assessment technique that employs an analytical hierarchy process. Due to the significant nonlinearity present in the closed loop bioprocess with substrate inhibition, the feedforward artificial neural network controller is trained using a closed-loop dataset, and its performances are compared with the conventional controllers. The controller has demonstrated its ability to manage realistic feed fluctuations without risking upset to the culture. The biomass concentration showed only minor deviations, settling swiftly back to the desired value by smoothly adjusting the dilution rate. This controller with tansig and purelin functions overcomes nonlinearities and time delays better than conventional controllers. The results suggest that the artificial neural network controller offers the desired simplicity and effectiveness for industrial applications.</div></div>","PeriodicalId":48907,"journal":{"name":"Journal of Computational Science","volume":"91 ","pages":"Article 102688"},"PeriodicalIF":3.1,"publicationDate":"2025-07-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144711668","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-07-21DOI: 10.1016/j.jocs.2025.102673
Joginder Singh , Shubhra Sankar Ray
MicroRNAs (miRNAs) are key biomarkers in cancer diagnosis and treatment. Identification of drug-resistant miRNAs may help in effective treatment of cancer. Two new z score based fuzzy rough relevance and redundancy entropies are developed and then a weighted framework is introduced to integrate the entropies for ranking and selecting miRNAs in classifying control and drug resistant patients. Here, two key components of soft computing, fuzzy set and rough set are utilized. The methodology is called a weighted framework for integrating fuzzy rough set-based relevance and redundancy entropies (WFIFRRRE). The z score is used to compute the fuzzy membership of expression values required for both entropies. Fuzziness deals with the overlapping nature of miRNA expression profiles and rough set helps in determining the exact class size. The weights in WFIFRRRE, assigned to relevance and redundancy entropies, are determined in a supervised manner to maximize the score used for validating the classification performance in discriminating the control and drug-resistant patients. The weights are varied from 0 to 1 in steps of 0.01 which enables an integration between relevance and redundancy entropies. A subset of miRNAs is selected from the ranked list and the performance is evaluated using three benchmark classifiers on eight drug-resistant cancer datasets. Experimental results show that WFIFRRRE provides better prediction accuracy than the popular methods used for comparison. The classification accuracy in terms of score, achieved by WFIFRRRE, ranges from 0.74 to 1.0, 0.75 to 1.0, and 0.73 to 1.0 using random forest, Naive Bayes, and linear SVM classifiers, respectively. The resultant set of miRNAs obtained using WFIFRRRE is also verified with the help of existing biological studies. The source code of WFIFRRRE is available at https://www.isical.ac.in/ shubhra/WFIFRRRE.html.
{"title":"Integrating fuzzy rough set-based entropies for identifying drug-resistant miRNAs in cancer","authors":"Joginder Singh , Shubhra Sankar Ray","doi":"10.1016/j.jocs.2025.102673","DOIUrl":"10.1016/j.jocs.2025.102673","url":null,"abstract":"<div><div>MicroRNAs (miRNAs) are key biomarkers in cancer diagnosis and treatment. Identification of drug-resistant miRNAs may help in effective treatment of cancer. Two new z score based fuzzy rough relevance and redundancy entropies are developed and then a weighted framework is introduced to integrate the entropies for ranking and selecting miRNAs in classifying control and drug resistant patients. Here, two key components of soft computing, fuzzy set and rough set are utilized. The methodology is called a weighted framework for integrating fuzzy rough set-based relevance and redundancy entropies (WFIFRRRE). The z score is used to compute the fuzzy membership of expression values required for both entropies. Fuzziness deals with the overlapping nature of miRNA expression profiles and rough set helps in determining the exact class size. The weights in WFIFRRRE, assigned to relevance and redundancy entropies, are determined in a supervised manner to maximize the <span><math><mi>F</mi></math></span> score used for validating the classification performance in discriminating the control and drug-resistant patients. The weights are varied from 0 to 1 in steps of 0.01 which enables an integration between relevance and redundancy entropies. A subset of miRNAs is selected from the ranked list and the performance is evaluated using three benchmark classifiers on eight drug-resistant cancer datasets. Experimental results show that WFIFRRRE provides better prediction accuracy than the popular methods used for comparison. The classification accuracy in terms of <span><math><mi>F</mi></math></span> score, achieved by WFIFRRRE, ranges from 0.74 to 1.0, 0.75 to 1.0, and 0.73 to 1.0 using random forest, Naive Bayes, and linear SVM classifiers, respectively. The resultant set of miRNAs obtained using WFIFRRRE is also verified with the help of existing biological studies. The source code of WFIFRRRE is available at <span><span>https://www.isical.ac.in/ shubhra/WFIFRRRE.html</span><svg><path></path></svg></span>.</div></div>","PeriodicalId":48907,"journal":{"name":"Journal of Computational Science","volume":"91 ","pages":"Article 102673"},"PeriodicalIF":3.7,"publicationDate":"2025-07-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144724960","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
扫码关注我们
求助内容：
应助结果提醒方式：
京公网安备 11010802042870号