A hybrid Constraint Programming/Mixed Integer Programming framework for the preventive signaling maintenance crew scheduling problem

Most farm work planning for land-use crops such as sugarcane belongs to flexible flow shop scheduling if neglecting cooperative work and other specific constraints. Because the conventional approaches to the flexible flow shop scheduling cannot formulate these specific constraints, we require a new approach for solving land-use crop planning problems that considers cooperative work. This paper describes a detailed mathematical model and a hybrid algorithm for solving the model, in which many practical constraints are taken into account, including cooperative work, optimum time windows, waiting time between operations, and moving time. The hybrid algorithm uses meta-heuristic simulated annealing and a mixed integer programming solver in Gurobi. In order to obtain good schedules in a reasonable time, we adopt a strategy of fixing partial work sequences in the simulated annealing procedure and optimizing the partial schedule using the solver. The results of the evaluation computation show that the proposed model is operative for the practical constraints, and that the hybrid algorithm is adaptable to scheduling computation. The strategy of fixing partial work sequences is applicable to reducing computation times for large-sized land-use crop planning problems.

The article provides a comprehensive analysis of strategies, technologies, and organizational models for the maintenance and repair (MRO) of thermal processing equipment within the context of continuous production cycles typical for metallurgical, chemical, and energy industries. The features of continuous production (straight-line flow, system interdependence, powerful substance and energy flows, high downtime cost, territorial localization) imposing specific requirements on the MRO system are considered. A comparative analysis of classic organizational approaches — preventive maintenance, condition-based maintenance, and run-to-failure maintenance — is conducted, justifying the necessity for a flexible, adaptive hybrid model suited for continuous processes. Modern diagnostic methods applied for predictive maintenance and specifics of repairing critical units of thermal equipment under strict time constraints of planned shutdowns are detailed. Special focus is placed on integrating digital technologies (digital twins, IIoT, AI) as tools for synchronizing material and maintenance flows, minimizing downtime, and maximizing the efficiency of shutdown windows. It is concluded that the optimal MRO system for continuous production must be dynamic, data-driven, and combine principles of predictive maintenance with rigorous scheduling to ensure process continuity and economic efficiency.

Optimal power flow is a widely used tool in power system planning and management. Due to the complexity of the power system both in terms of number of variables, degrees of freedom and uncertainty, there is a continuous effort to find more efficient computational methods to solve optimal power flow problems. This article presents a novel method to speed-up the solution of a security constraint optimal power flow problem. An unconventional warm start based on the training of a neural network is investigated as an option to improve the computational efficiency of the optimization problem. The principle of the method and the validity of the approach is demonstrated by different analysis performed on the IEEE14 test grid and based on a linearized mathematical formulation of the problem. The results show the effectiveness of the method in reducing the number of iterations needed to converge to global optimum.

The problem we consider is short‐term harvesting planning for a total planning period of 4–6 weeks where we want to decide the harvest sequences or schedules for harvest crews. A schedule is an order or sequence of harvest areas assigned to each crew. The harvesting of areas is planned in order to meet industrial demand. The total cost includes harvesting, transportation, and storage. One considerable cost is due to the quality reduction of logs stored at harvest areas. There are a number of restrictions to be considered. Areas are of varying size and the composition of assortments in each area is different. Each harvest team has different skills, a different home base, and different production capacity. Another aspect is the road network. There is a cost related to road opening (restoring, snow removal). In this paper, we develop a mixed integer programming (MIP) model for the problem. The schedules are represented by 0/1 variables. With a limited number of schedules, the problem can be solved by a commercial MIP solver. We have also developed a heuristic solution approach that provides high‐quality integer solutions within a distinct time limit to be used when more schedules are used. Computational results from a major Swedish forest company are presented.

Self-growth learning-based machine scheduler to minimize setup time and tardiness in OLED display semiconductor manufacturing

The accurate prognostics for actuator malfunctions is a challenging task. Developing reliable prognostics methods is vital for providing reasonable preventive maintenance. Particle filter has been proved to be a prevailing approach to cope with actuator prognostics problems. However, the measurement function in the particle filter algorithm cannot be obtained in the prediction process. To this end, this article presents a hybrid framework combining support vector regression and particle filter. To accomplish the accurate prognostics for actuator fault of civil aircraft and provide the reliable “measurements” for the subsequent particle filter algorithm, the traditional support vector regression algorithm needs to be improved, and the error confidence level is imported to evaluate the usability of the support vector regression prediction output quantitatively. In addition, an improved particle filter based on Kendall correlation coefficient is put forward to address the problem of particles’ degeneracy. The experimental results are presented, demonstrating that the support vector regression–particle filter hybrid framework has satisfactory performance with better prognostics accuracy and higher fault resolution than traditional approaches.

The early detection of lung diseases, including cancer, is essential for improving patient outcomes. However, traditional diagnostic approaches and standard deep learning models often face challenges in effectively analyzing the complex spatial and temporal variations in medical imaging data, particularly in CT scans. To address these challenges, we propose DC-CAD, a novel hybrid framework that integrates Dilated Capsule Networks, Channel-wise Attention Mechanisms, and Distanced Long Short-Term Memory for precise and early diagnosis of lung diseases. DC-CAD is innovative in its ability to combine multi-scale feature extraction and temporal dynamic analysis, enabling the model to capture intricate spatial relationships and sequential changes in lung tissue. The model consists of three main contributions: (1) Dilated Capsule Networks for improved multi-scale context aggregation, which captures subtle textural variations, (2) a Channel-wise Attention Mechanism to focus on the most relevant regions of interest, minimizing the impact of irrelevant features, and (3) Distanced LSTM layers to model temporal dependencies across sequential CT scans, providing insights into disease progression. Through comprehensive experiments on the LC25000 dataset, DC-CAD achieves 99.52% accuracy, significantly outperforming baseline models such as standard Capsule Networks and Convolutional Neural Networks. The model also reduces the error rate to 0.48%, demonstrating substantial improvements in diagnostic performance, including increased accuracy, sensitivity, and specificity. These results establish DC-CAD as a powerful and reliable tool for automated lung disease diagnosis, with significant potential to enhance clinical workflow by reducing radiologists’ workload through its interpretability and efficiency. Moving forward, we plan to extend the model to handle multi-modal data and investigate advanced attention mechanisms to further improve diagnostic accuracy and generalizability.

The adoption of wind energy is pivotal for advancing sustainable power systems, particularly in off-grid microgrids where infrastructure limitations hinder conventional energy solutions. The inherent variability of wind generation, however, challenges grid reliability and demand–supply balance, necessitating accurate forecasting models. This study proposes a hybrid framework for short-term wind speed prediction, integrating deep learning (Long Short-Term Memory, LSTM) and ensemble methods (random forest, Extra Trees) to exploit their complementary strengths in modeling temporal dependencies. A multivariate approach is adopted using meteorological data (including wind speed, temperature, humidity, and pressure) to capture complex weather interactions through a structured time-series design. The framework also includes a feature selection stage to identify the most relevant predictors and a hyperparameter optimization process to improve model generalization. Three wind speed variables, maximum, average, and minimum, are forecasted independently to reflect intra-day variability and enhance practical usability. Validated with real-world data from Cuenca, Ecuador, the LSTM model achieves superior accuracy across all targets, demonstrating robust performance for real-world deployment. Comparative results highlight its advantage over tree-based ensemble techniques, offering actionable strategies to optimize wind energy integration, enhance grid stability, and streamline renewable resource management. These insights support the development of resilient energy systems in regions reliant on sustainable microgrid solutions.

This paper introduces a novel hybrid framework, ICEEMDAN-PatchTST+FDPO, which integrates advanced signal decomposition, Transformer-based time series forecasting, and forecast-driven portfolio optimization (FDPO). Improved Complete Ensemble Empirical Mode Decomposition with Adaptive Noise (ICEEMDAN) enhances signal stationarity by decomposing noisy financial series into intrinsic mode functions, while PatchTST employs patch-wise self-attention to capture complex temporal dependencies. These components enable precise multi-step return forecasting, which is subsequently operationalized through FDPO to generate dynamically optimized allocations aligned with predicted market movements. Extensive out-of-sample evaluations from January 2022 to December 2024 on a rule-based selection of ten representative NIFTY 50 constituents demonstrate that the proposed framework delivers substantial gains over conventional baselines (EqualWeight, Mean–Variance Optimization, Risk Parity) and alternative deep learning models (LSTM, TimesNet, NBEATSx, Informer) in both forecasting accuracy and portfolio performance. The model achieves the lowest MAE (38.25), RMSE (55.92), and MAPE (2.06%), with an <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">R</i><sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sup> of 0.9776. It also attains a CAGR of 1.8453, a Sharpe ratio of 3.615, a Sortino ratio of 3.985, and the lowest maximum drawdown of –0.0971. Alignment metrics such as Cosine Similarity and Jensen–Shannon Divergence further indicate that its weight vectors closely approximate ideal allocations based on actual future returns. A supplementary validation on a rule-based S&P 500 subset corroborates these trends, suggesting that the framework maintains its performance advantages across distinct market environments. Overall, the findings highlight the practical viability of the ICEEMDAN-PatchTST+FDPO framework for robust and adaptive asset management under real-world market dynamics.

3D human pose estimation (3DHPE) has evolved into a sophisticated and pivotal technique, emerging as a prominent research focus in computer vision and robotics, based on the power of deep neural networks. Unlike conventional image-based methods, the integration of video data introduces rich temporal information, enabling the exploitation of spatiotemporal correlations to achieve more accurate and robust results. However, modeling both local motion continuity and long-range temporal dependencies remains a significant challenge, as recurrent architectures such as LSTMs excel at capturing local dynamics but struggle with long-term information retention, whereas Transformer-based models provide strong global reasoning yet are less sensitive to fine-grained motion details. In this research, a hybrid LSTM–Transformer network that effectively integrates the local feature sensitivity of LSTMs with the global coordination ability of Transformers was proposed to cover these deficiencies. Multiple fusion strategies were investigated to evaluate their impact on 3DHPE performance. Comprehensive experiments finished on benchmark datasets demonstrate that the proposed Single-branch LSTM+Transformer architecture achieves the most competitive results, yielding a mean per-joint position error (MPJPE) of 49.51 mm, which outperforms conventional Transformer- based-only and LSTM-based-only models. This hybrid framework provides a novel and effective paradigm for future research in video-based human motion understanding.

Operating room (OR) scheduling is a critical challenge in healthcare, directly impacting patient outcomes and hospital efficiency. Traditional methods often struggle with the complex, multi-level constraints and uncertainties inherent in OR scheduling, such as resource limitations, variable surgery durations, and emergency cases. This study aims to develop a novel hybrid framework that optimizes OR scheduling by integrating multi-level optimization with reinforcement learning and column generation techniques. The proposed framework decomposes the OR scheduling problem into strategic, tactical, and operational levels, enabling focused optimization at each layer while ensuring cohesive decision-making across the hierarchy. Reinforcement learning guides the column generation process, learning policies that identify promising scheduling options to enhance solution quality and computational efficiency. Robust uncertainty handling mechanisms are incorporated to manage variability in surgery durations and resource availability without compromising tractability. Experiments were conducted using three years of real-world data from Shanxi Provincial People's Hospital, complemented by large-scale synthetic datasets to evaluate scalability and robustness of the framework. The framework demonstrates meaningful improvements in key operational metrics compared to traditional approaches. Analysis of three years of implementation shows consistent enhancements in operational efficiency, including a reduction in average patient waiting time by 15.8% (from 10.1 to 8.5 days), an increase in OR utilization by 5.4 percentage points (from 73.8% to 79.2%), and improved workload balance among surgeons. The framework maintains robust performance under uncertainty, achieving a 92.5% feasibility rate and reducing schedule disruptions by 26.2%. The proposed hybrid framework offers a practical and scalable solution for optimizing OR scheduling, demonstrating improvements in healthcare delivery and operational performance in real hospital environments. By effectively balancing multiple operational objectives while handling practical constraints and uncertainties, the framework provides a viable approach for healthcare systems seeking incremental yet sustainable improvements in efficiency and patient care.

Real-world entities' behaviors, associated with their side information, are often recorded over time as asynchronous event sequences. Such event sequences are the basis of many practical applications, neural spiking train study, earth quack prediction, crime analysis, infectious disease diffusion forecasting, condition-based preventative maintenance, information retrieval and behavior-based network analysis and services, etc. Temporal point process (TPP) is a principled mathematical tool for the modeling and learning of asynchronous event sequences, which captures the instantaneous happening rate of the events and the temporal dependency between historical and current events. TPP provides us with an interpretable model to describe the generative mechanism of event sequences, which is beneficial for event prediction and causality analysis. Recently, it has been shown that TPP has potentials to many machine learning and data science applications and can be combined with other cutting-edge machine learning techniques like deep learning, reinforcement learning, adversarial learning, and so on. We will start with an elementary introduction of TPP model, including the basic concepts of the model, the simulation method of event sequences; in the second part of the tutorial, we will introduce typical TPP models and their traditional learning methods; in the third part of the tutorial, we will discuss the recent progress on the modeling and learning of TPP, including neural network-based TPP models, generative adversarial networks (GANs) for TPP, and deep reinforcement learning of TPP. We will further talk about the practical application of TPP, including useful data augmentation methods for learning from imperfect observations, typical applications and examples like healthcare and industry maintenance, and existing open source toolboxes.

This work presents a hybrid approach to solve the maximum stable set problem, using constraint and semidefinite programming. The approach consists of two steps: subproblem generation and subproblem solution. First we rank the variable domain values, based on the solution of a semidefinite relaxation. Using this ranking, we generate the most promising subproblems first, by exploring a search tree using a limited discrepancy strategy. Then the subproblems are being solved using a constraint programming solver. To strengthen the semidefinite relaxation, we propose to infer additional constraints from the discrepancy structure. Computational results show that the semidefinite relaxation is very informative, since solutions of good quality are found in the first subproblems, or optimality is proven immediately.KeywordsCombinatorial Optimization ProblemConstraint ProgrammingLinear RelaxationMaximum Clique ProblemSemidefinite RelaxationThese keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.

Supply chain demand forecasting is a core component of modern enterprise operations management, directly impacting inventory optimization, production planning, and customer satisfaction. Traditional statistical methods such as Autoregressive Integrated Moving Average (ARIMA) models face limitations when handling complex nonlinear patterns and multidimensional features. While single machine learning models enhance prediction accuracy, they struggle to simultaneously capture structured features and temporal dependencies.This study proposes an innovative hybrid forecasting framework integrating Extreme Gradient Boosting (XGBoost) with Long Short-Term Memory (LSTM) networks. XGBoost excels at learning high-dimensional structured features, while LSTM specializes in modeling temporal dependency patterns. An adaptive weight fusion mechanism enables complementary strengths between the two models.Empirical analysis using the Kaggle open-source retail dataset (encompassing 10 stores, 50 products, and 913,000 transaction records) demonstrates that the hybrid framework significantly outperforms both single models and traditional methods across multiple metrics, including Mean Absolute Percentage Error (MAPE) and Root Mean Square Error (RMSE).The findings provide supply chain managers with actionable forecasting tools, offering significant theoretical and practical value for enhancing prediction accuracy and operational efficiency.

Bearings are important elements of mechanical systems and the correct forecasting of their remaining useful life (RUL) is key to successful predictive maintenance. Nevertheless, noise interference during different operating conditions is also a significant problem in predicting their RUL. Existing denoising-based RUL prediction models often show degraded performance when exposed to heterogeneous and non-stationary noise, resulting in unstable feature extraction and reduced generalisation. To address the challenge of heterogeneous and non-stationary noise in bearing RUL prediction, this study proposes a hybrid framework that combines a noise-conditioned convolutional denoising autoencoder (NC-CDAE) and a temporal attention transformer (TAT). The NC-CDAE adaptively suppresses diverse noise types through conditional modulation, while the TAT captures long-term temporal dependencies to enhance degradation trend learning. This synergistic design improves both the noise robustness and temporal modelling capability of the system. To further validate the model under varying conditions, synthetic datasets with different noise intensities were generated using a conditional generative adversarial network (cGAN). Comprehensive experiments show that the proposed NC-CDAE + TAT framework achieves lower and more stable errors than state-of-the-art methods, reducing RMSE by up to 23.6% and MAE by 18.2% on average and maintaining consistent performance (an RMSE between 0.155 and 0.194) across diverse conditions.