International Journal of Computing Science and Mathematics (11 papers in press)

Regular Issues

• Artistic Mural Reconstruction via a GAN Network based on Sorting Loss  
  by Zhiqiang Chen, Jiaqi Liu, Cao Jianfang, Cunhe Peng 
  Abstract: To address severe mural image defects and the low-resolution and rough reconstruction details of mural reconstruction methods, which can lead to a reduction in mural artistry. This paper presents an artistic reconstruction method (RSSRGAN) for murals. This method adopts the architecture of generative adversarial networks and introduces an attention mechanism. First, channel separation is performed on the feature map obtained during preliminary feature extraction, and weight prediction is performed on the 64-dimensional features to construct the channel dependence between the mural feature maps to retain the high-frequency features lost by the murals in the LR space. Finally, mural textural features are retained to improve the artistic reconstruction effect. Compared with the four popular superresolution reconstruction baseline models, the proposed method achieves a peak signal-to-noise ratio (PSNR) increase of more than 0.58 and an increase in the structural similarity index (SSIM) of more than 0.025 on the mural dataset. Moreover, public dataset verification on the DIV2K dataset showed that the method achieved good reconstruction quality, in which the PSNR increased by more than 0.27 and the SSIM increased by more than 0.014. The RSSRGAN method has achieved significant improvements in mural image reconstruction and provides a new and effective method for artistic mural reconstruction.
  Keywords: mural protection; superresolution reconstruction; generative adversarial network; twin neural network; attention mechanism.
  DOI: 10.1504/IJCSM.2025.10075071
   
  
• Thermal Convection of a Oldroyd-B Nanofluid with Coriolis Effect  
  by Abhishek Singh, Mala Mala 
  Abstract: The investigation focuses on the onset of convection in a horizontal layer with the inclusion of Oldroyd-B nanofluid. The non-dimensional governing equations is solved using the normal mode technique, resulting in an eigenvalue problem. Analytical expression for Rayleigh number is obtained. Critical Rayleigh number values are determined for specific parameter settings. The influence of dimensionless parameters such as the Lewis number (Le), Prandtl number (Pr), Modified particle density increment (Nb), modified diffusivity ratio (Na), Taylor number (Ta), Nanoparticle Rayleigh number (Rn), and the relaxation times of the fluid 1 and 2 on the critical Rayleigh number is analysed. The results of the study indicate that the Taylor number (Ta) acts as a stabilising factor for the system, while the modified diffusivity ratio (Na) and Nanoparticle Rayleigh number (Rn) function as destabilising factors. Moreover, the critical Rayleigh number exhibits a non-monotonic dependency on the coefficients 1 and 2.
  Keywords: Oldroyd-B nanofluid; Linear stability analysis; Thermal convection.
  DOI: 10.1504/IJCSM.2025.10075311
   
  
• Modeling the Impacting of Extreme Snow and Ice Conditions on Energy Flow in Integrated Energy Systems  
  by Jian Wang, Hao Li, Zhanxi Zhang, Jieshan Shan, Fu Shen, Hongchun Shu, Yiming Han, Zilong Cai, Kaizheng Wang, Lei Kou 
  Abstract: This paper proposes a novel framework for analysing the dynamic effects of extreme snow and ice (S&I) conditions on energy flow within integrated energy systems (IES). First, an enhanced IES framework is developed to strengthen the coupling between electricity, heat, gas systems, energy hubs, and renewable energy sources. Second, an improved transmission line icing failure model is introduced, considering the variations in icing thickness and the breaking force. Monte Carlo simulations and numerical calculations are applied to assess the impacts of extreme S&I events on energy flow in IES. A 64-bus case study illustrates the significant operational differences and safety risks arising from such extreme weather conditions. The proposed IES framework provides a comprehensive view of complex energy interactions and underscores the need for resilient systems.
  Keywords: Integrated energy systems; dynamic energy flow; icing; transmission line ice failure.
  DOI: 10.1504/IJCSM.2025.10075665
   
  
• Deep Reinforcement Learning for Dynamic Cellular Manufacturing Systems with Deterioration Effect  
  by Mostafa Jafari, Amir Hossein Akbari 
  Abstract: This study presents an integrated framework for optimising machine layout and production planning in dynamic cellular manufacturing systems under uncertainty. The framework addresses key challenges including machine deterioration and breakdowns, order rejection, and tardiness costs, which are often treated separately in traditional approaches. A multi-objective mathematical model is developed to maximise profit, increase the number of accepted orders, and balance machine workloads to reduce failures. The solution employs a three-step hierarchical approach: heuristic machine-to-cell assignment, deep reinforcement learning for real-time order acceptance and scheduling while considering machine deterioration, and heuristic layout refinement. Computational results show that the proposed method accepts 2.63% more orders with a 5.7% profit reduction, enhancing customer attraction and competitiveness. Workload balancing decreases machine repairs by 11.5%, improving system stability and reducing maintenance costs. Despite an average profit loss of 9.77% due to machine deterioration, the framework significantly improves efficiency and operational resilience in dynamic manufacturing environments.
  Keywords: Cellular manufacturing System; Orde Acceptance and Scheduling; Deterioration effect; deep reinforcement learning.
  DOI: 10.1504/IJCSM.2025.10075736
   
  
• Multi-Stage Adaptive Firefly Algorithm with Enhanced Search  
  by Kefeng Li, Na Jin, Tang Jun, Cao Yiqing 
  Abstract: As a popular swarm intelligence optimization approach, firefly algorithm (FA) has exhibited excellent search capabilities in various optimization problems. However, FA still has some limitations. The search efficiency is sensitive to the step size factor, and the single search pattern results in slow convergence rate. To tackle these issues, this paper proposes a multi-stage adaptive FA with enhanced search (namely MSAFAES). First, a new adaptive parameter method is designed, in which the entire search is divided into two stages. Different parameters strategies are adopted for different search stages. Then, three types of search patterns are employed in the search process. To validate the performance of MSAFAES, ten well-known benchmark problems are tested. Computational results demonstrate the effectiveness of MSAFAES when compared with three other FA variants.
  Keywords: Firefly algorithm; multi-stage; adaptive; multi-strategy; optimization.
  DOI: 10.1504/IJCSM.2025.10075779
   
  
• Adaptative strategies for solving partial differential equations by Kansa's method  
  by Selma Bouzit 
  Abstract: This work presents an innovative and efficient meshless method for solving high-dimensional partial differential equations (PDEs). By utilising generalised multiquadric radial basis functions (GMQRBF) with an exponent β, the method incorporates various shape parameter c selection strategies to enhance numerical accuracy. Three approaches: optimal, trigonometric, and random for β and c are analysed for their performance across different problems. The method's mathematical foundation is rigorously studied, and extensive numerical experiments confirm its accuracy and robustness in solving linear and nonlinear PDEs across various dimensions. The results demonstrate its potential as a reliable and versatile tool for high-dimensional PDE applications.
  Keywords: PDEs; partial differential equations; linear problems; non-linear problems; RBFs; radial basis functions; Kansa method; shape parameter.
  DOI: 10.1504/IJCSM.2025.10074451
   
  
• Learning-assisted empirical mode decomposition algorithm for pre-stack seismic wave impedance prediction  
  by Shu-cheng Sun, Bo Tang, Xue-song Yan 
  Abstract: Among numerous oil and gas exploration technologies, pre-stack seismic inversion technology based on reservoir elastic parameter information that can reflect more stratigraphic characteristics has become a popular technology in seismic inversion due to its ability to improve the accuracy of exploration. This paper proposes a learning-assisted empirical mode decomposition (EMD) algorithm. This algorithm addresses the problems of over envelope and under envelope in EMD algorithms, using segmented cubic Hermite interpolation algorithm instead of cubic spline interpolation, and using feature scale extension to reduce endpoint influence. To eliminate invalid components, correlation coefficients are used to remove some invalid components and reduce modal aliasing. In order to verify the performance of the proposed algorithm, traditional data processing methods and improved methods were compared with other algorithms to verify the improvement of wave impedance parameter prediction accuracy of the proposed algorithm.
  Keywords: pre-stack seismic inversion; elastic parameters; EMD; empirical mode decomposition; CNN; convolutional neural network.
  DOI: 10.1504/IJCSM.2025.10074359
   
  
• Global-local temporal attention network for short-term solar irradiance prediction  
  by Jian Wang, Huiyuan Liu, Zhichong Dou, Jieshan Shan, Zhanxi Zhang, Fu Shen, Hongchun Shu, Yiming Han, Zilong Cai, Kaizheng Wang, Meng Wang, Dongkai Zhang, Lei Kou, Huiyuan Nie 
  Abstract: The intermittent and fluctuating nature of solar radiation poses significant challenges to power system stability, as it can lead to unpredictable fluctuations in energy generation, thereby complicating grid management. To address this, we propose a solar irradiance prediction method based on the global-local temporal attention network (GLTAN). The GLTAN comprises three key components: the global-local feature extraction (GLFE) module, the temporal attention mechanism (TAM), and the gated recurrent unit (GRU). The GLFE module uses a dual-layer transformer and temporal convolutional network (TCN) to extract both global and local features, capturing short-term and long-term trends. The TAM selectively highlights relevant information, while the GRU captures short-term dependencies. Experimental results for four different prediction time steps show that GLTAN outperforms the other models, with an average of 7.3% improvement in R2 over nine prediction steps.
  Keywords: solar irradiance; short-term prediction; GLFE; global-local feature extraction; self-attention mechanism.
  DOI: 10.1504/IJCSM.2025.10074450
   
  
• Simulation optimisation of Maicai delivery location using the elbow method and K-means clustering algorithm  
  by Yonghua Lu, Bingda Zhang, Yuting Shen, Zixia Chen 
  Abstract: This study proposes an innovative integration of the elbow method and the K-means clustering algorithm to optimise delivery centre locations for Maicai. Based on 55 demand sites, a model is constructed to analyse geographic positions and order volumes. The elbow method is first applied to determine the optimal number of clusters (k), followed by K-means clustering to iteratively optimise delivery centre locations. Python is used to generate a k-S.S.E relationship curve and identify the inflection point automatically. This approach provides data-driven support for determining the optimal number of clusters and uses K-means to perform spatial clustering of demand points. Empirical results demonstrate that the hybrid method effectively reduces logistics costs and enhances delivery efficiency, demonstrating operational efficiency improvements for real-world delivery operations.
  Keywords: delivery-centre location; elbow method; K-means clustering; logistics optimisation; delivery efficiency.
  DOI: 10.1504/IJCSM.2025.10074316
   
  
• An intelligent method for human activity recognition based on feature fusion of point clouds from FMCW millimetre-wave radar  
  by Yingbo Wu, Dunyu Zhao 
  Abstract: To achieve efficient and accurate recognition of human activities, this paper proposes a human activity recognition (HAR) method based on point cloud feature fusion of millimeter-wave radar. To fully extract the features of the activity targets, a point-cloud feature fusion model is designed that combines convolutional neural network (CNN) and attention-based dynamic graph convolutional neural network (DGCNN). CNN captures Doppler features from the Micro-Doppler Spectrogram, while the DGCNN extracts key information from the spatio-temporal features of the point cloud. The combination of these two networks enables activity recognition. Experiments have validated the effectiveness and accuracy of the recognition method.
  Keywords: millimeter-wave radar; radar signal processing; activity recognition; neural network.
  DOI: 10.1504/IJCSM.2025.10075116
   
  
• An efficient hash-based assessment and recovery algorithm for distributed healthcare systems  
  by Ramzi A. Haraty, Mohamad Jaber 
  Abstract: Advancements in healthcare information technology have improved data sharing and patient care but also introduced cybersecurity risks. This study proposes a distributed algorithm to assess damage from cyberattacks and accelerate database recovery. The algorithm influences hash tables for efficient identification and retrieval of affected transactions, minimising execution time and resource consumption. A checkpoint mechanism further optimises performance by discarding outdated logs. Comparative analysis demonstrates the algorithm's effectiveness in mitigating attack impact and enhancing healthcare system resilience. By integrating hash tables for rapid data retrieval, this approach offers a robust solution to safeguard critical medical data.
  Keywords: damage assessment; database recovery; distributed databases; information warfare; healthcare systems.
  DOI: 10.1504/IJCSM.2025.10074482