Manufacturers are increasingly adopting manufacturing data analytics (MDA) as a key factor for smart manufacturing. However, successful MDA implementation remains limited due to various issues. Existing studies have barely suggested a comprehensive such issues by focusing on parts of technological, organizational, and environmental (TOE) contexts or issues limited to partial steps of MDA. This study addresses these gaps by identifying comprehensive issue set for MDA implementation (CISM) through a systematic review of 35 papers. The 29 distinct issues with 9 categories were derived to cover both TOE contexts and the five major steps of MDA. The comprehensiveness of CISM was validated through three real-world MDA implementation case studies. CISM is expected to suggest issues for manufacturers to address proactively in MDA implementation and to serve as a basis for stimulating future studies on MDA implementation. • Manufacturers are increasingly attempting to implement manufacturing data analytics (MDA). • Existing studies rarely offer a comprehensive set of issues for MDA implementation. • This study constructs a comprehensive set of 29 issues for MDA implementation across TOE contexts and the 5 major MDA steps • CISM utility was validated with three real-world MDA case studies. • CISM will support manufacturers identify and address key issues in implementing data-driven cyber manufacturing.

Time-series forecasting (TSF) is a traditional problem in the field of artificial intelligence, and models such as recurrent neural network, long short-term memory, and gate recurrent units have contributed to improving its predictive accuracy. Furthermore, model structures have been proposed to combine time-series decomposition methods such as seasonal-trend decomposition using LOESS. However, this approach is learned in an independent model for each component, and therefore, it cannot learn the relationships between the time-series components. In this study, we propose a new neural architecture called a correlation recurrent unit (CRU) that can perform time-series decomposition within a neural cell and learn correlations (autocorrelation and correlation) between each decomposition component. The proposed neural architecture was evaluated through comparative experiments with previous studies using four univariate and four multivariate time-series datasets. The results showed that long- and short-term predictive performance was improved by more than 10%. The experimental results indicate that the proposed CRU is an excellent method for TSF problems compared to other neural architectures.

Manufacturers are increasingly adopting manufacturing data analytics (MDA) as a key factor for smart manufacturing. However, successful MDA implementation remains limited due to various issues. Existing studies have barely suggested a comprehensive such issues by focusing on parts of technological, organizational, and environmental (TOE) contexts or issues limited to partial steps of MDA. This study addresses these gaps by identifying comprehensive issue set for MDA implementation (CISM) through a systematic review of 35 papers. The 29 distinct issues with 9 categories were derived to cover both TOE contexts and the five major steps of MDA. The comprehensiveness of CISM was validated through three real-world MDA implementation case studies. CISM is expected to suggest issues for manufacturers to address proactively in MDA implementation and to serve as a basis for stimulating future studies on MDA implementation. • Manufacturers are increasingly attempting to implement manufacturing data analytics (MDA). • Existing studies rarely offer a comprehensive set of issues for MDA implementation. • This study constructs a comprehensive set of 29 issues for MDA implementation across TOE contexts and the 5 major MDA steps • CISM utility was validated with three real-world MDA case studies. • CISM will support manufacturers identify and address key issues in implementing data-driven cyber manufacturing.

Time-series forecasting (TSF) is a traditional problem in the field of artificial intelligence, and models such as recurrent neural network, long short-term memory, and gate recurrent units have contributed to improving its predictive accuracy. Furthermore, model structures have been proposed to combine time-series decomposition methods such as seasonal-trend decomposition using LOESS. However, this approach is learned in an independent model for each component, and therefore, it cannot learn the relationships between the time-series components. In this study, we propose a new neural architecture called a correlation recurrent unit (CRU) that can perform time-series decomposition within a neural cell and learn correlations (autocorrelation and correlation) between each decomposition component. The proposed neural architecture was evaluated through comparative experiments with previous studies using four univariate and four multivariate time-series datasets. The results showed that long- and short-term predictive performance was improved by more than 10%. The experimental results indicate that the proposed CRU is an excellent method for TSF problems compared to other neural architectures.

Changes in recent business environments have created the necessity for a more efficient and effective business process management. The workflow management system is software that assists in defining business processes as well as automatically controlling the execution of the processes. We propose a new approach to the automatic execution of business processes using event-condition-action (ECA) rules that can be automatically triggered by an active database. First of all, we propose the concept of blocks that can classify process flows into several patterns. A block is a minimal unit that can specify the behaviors represented in a process model. An algorithm is developed to detect blocks from a process definition network and transform it into a hierarchical tree model. The behaviors in each block type are modeled using ACTA formalism. This provides a theoretical basis from which ECA rules are identified. The proposed ECA rule-based approach shows that it is possible to execute the workflow using the active capability of database without users' intervention. The operation of the proposed methods is illustrated through an example process.

The radar detection weapon system is a surveillance and reconnaissance system designed to precisely detect long-range targets and determine their position and status. A critical component of the radar detection weapon system is the wave guide, which serves as a path for high-frequency signals. Maintaining appropriate temperature and humidity levels is essential to ensure seamless signal transmission, and this is regulated using an air compressor. However, if compressed air leakage occurs, temperature and humidity may rise, leading to performance degradation and potential component failure, necessitating early fault detection. The sensor data collected form actual operations exhibits high-dimensional and nonlinear characteristics, with abnormal patterns appearing irregularly, making numerical approaches challenging. Furthermore, such methods face limitations in effectively learning overall patterns and dependencies over time. In this paper, we propose an anomaly detection model that combines time-domain feature with image encoding to effectively capture nonlinear signal patterns. Experimental result using real-world operational data demonstrate that the proposed method outperforms existing approaches. The proposed approach is expected to be useful for applying Condition-Based Maintenance(CBM) to various equipment that utilizes time-series data.

Predictive process monitoring (PPM) faces challenges when process patterns evolve over time, as prediction models experience performance degradation due to concept drift while losing accuracy on previously learned process variants through catastrophic forgetting. This study addresses these challenges in outcome-oriented PPM by integrating continual learning principles with process mining domain knowledge. We propose an adaptive memory mechanism that extracts process structure information through activity relationship analysis, categorizing relationships between activities as "Always Follows," "Sometimes Follows," or "Never Follows." The mechanism computes relation entropy vectors to characterize process patterns and uses cosine similarity between current and historical process patterns to adaptively weight memory buffers during model training. This allows the model to weight historically relevant process knowledge based on pattern similarity. We evaluate the mechanism using stability, plasticity, and trade-off metrics on the Business Process Intelligence Challenge 2015 dataset across multiple temporal splits. Results show that the adaptive memory mechanism achieves stability of 0.87 ± 0.06, plasticity of 0.84 ± 0.11, and trade-off of 0.85 ± 0.08, achieving competitive performance compared to established baseline methods. The mechanism provides explainability through visualizations that demonstrate how the model weights different memory buffers during training. Validation using concept drift detection techniques confirms that the mechanism’s behavior aligns with observable process changes, indicating adaptation to evolving process patterns while reducing catastrophic forgetting.