Abstract The paper proposes a deep learning based approach to enhance the efficiency, stability, and control of hybrid renewable energy systems. It addresses challenges such as the intermittency and variability of sources like solar, wind, and geothermal energy,... The paper proposes a deep learning based approach to enhance the efficiency, stability, and control of hybrid renewable energy systems. It addresses challenges such as the intermittency and variability of sources like solar, wind, and geothermal energy, which hinder optimal energy utilization and grid stability. The study introduces an LSTM based multi channel convolutional autoencoder model that effectively extracts temporal and spectral features from system signals, improving prediction accuracy by 9.8% over traditional CNN models. The architecture combines convolutional neural networks (CNNs) for noise reduction and feature extraction with long short term memory (LSTM) networks for capturing temporal dependencies. Experimental results using real photovoltaic data from Shaoxing, China (2014–2018) demonstrate that the Conv LSTM model achieves superior performance across key metrics MAAPE, RMSE, and MAE especially for forecasting beyond 15 minutes. The research also highlights the importance of integrating environmental variables, such as temperature, which shows a correlation coefficient of 0.5539 with solar output. By leveraging advanced deep learning techniques, the system enables adaptive control strategies, demand forecasting, and optimized energy scheduling, ultimately improving economic viability and operational reliability. Despite its promise, the approach faces challenges in data collection, model interpretability, and real time deployment, pointing to directions for future research. Read More Read Less

Abstract The paper proposes an integrated safety assessment algorithm for road construction by combining the Analytic Hierarchy Process (AHP) and Back Propagation (BP) Neural Network. Traditional safety evaluation methods, such as fuzzy comprehensive evaluation, lack adaptability to dynamic... The paper proposes an integrated safety assessment algorithm for road construction by combining the Analytic Hierarchy Process (AHP) and Back Propagation (BP) Neural Network. Traditional safety evaluation methods, such as fuzzy comprehensive evaluation, lack adaptability to dynamic environmental changes and struggle to update weights in real time. To address these limitations, the authors leverage AHP for systematic determination of indicator weights and the BP neural network for its strong nonlinear modeling, self learning, and adaptive capabilities. Four key safety indicators geological disasters, hazardous sources, hidden dangers, and compliance with safety standards are selected for evaluation. The AHP method quantifies their relative importance using expert scoring and pairwise comparisons. At the same time, the BP neural network processes input data through a three-layer architecture to compute a safety risk coefficient. This coefficient maps to one of five predefined safety risk levels, enabling practical risk categorization. Experimental validation on a real infrastructure project demonstrates that the integrated algorithm significantly outperforms conventional approaches, achieving an average relative error of only 0.004 compared to 0.0468 for traditional methods. Results confirm the model’s high accuracy, reliability, and effectiveness in predicting safety risks. Despite its advantages, the authors acknowledge remaining limitations and call for further refinement to better align with public safety expectations. The study contributes a robust, data driven framework for improving safety management in complex road construction environments. Read More Read Less

Abstract This paper explores enhancements to the Multi Objective Particle Swarm Optimization (MOPSO) algorithm to address challenges in educational optimization, particularly in intelligent and online teaching environments. It highlights the limitations of traditional teaching methods in applied universities... This paper explores enhancements to the Multi Objective Particle Swarm Optimization (MOPSO) algorithm to address challenges in educational optimization, particularly in intelligent and online teaching environments. It highlights the limitations of traditional teaching methods in applied universities and advocates integrating digital innovations such as MOOCs and flipped classrooms. The study focuses on improving MOPSO’s ability to handle constrained, unconstrained, and partially constrained multi objective problems by enhancing particle selection, maintaining solution diversity, and avoiding local optima. The proposed AAD MOPSO algorithm demonstrates superior performance compared to IMOPSO and PCCSMOPSO, especially in maintaining highquality external archives and achieving better convergence and distribution on benchmark ZDT test functions (ZDT1–ZDT6). Experimental results show that AAD MOPSO accurately approximates the true Pareto front with improved uniformity and coverage. The paper concludes that advanced computational intelligence algorithms can support more effective educational strategies by optimizing complex, multi faceted objectivessuch as fostering students’ moral development, critical thinking, and practical skills aligning with the needs of digital native learners. Ultimately, the research contributes a robust optimization framework applicable not only to education but also to other domains requiring efficient, multi objective decision making. Read More Read Less