A Management Science Framework for Predictive Optimisation Using the EDNN–LR Model: Algorithmic Insights from Industry 4.0 Systems

Artificial Intelligence (AI) is increasingly central to modern system engineering and service operations, enabling real-time decision-support in cyber-physical and data-intensive environments. This study develops an Extended Deep Neural Network–Logistic Regression (EDNN–LR) hybrid framework as a scalable AI solution for predictive optimisation within Industry 4.0 decision systems. The model integrates the nonlinear learning capability of deep neural networks with the interpretability and convergence stability of logistic regression, thereby enhancing transparency, robustness, and computational efficiency in engineering applications characterised by uncertainty and behavioural variability. The proposed framework is validated using a publicly available financial–cyber dataset comprising over 4.44 million records from CoinMarketCap (2013–2025), representing a dynamic cyber-physical decision environment analogous to complex industrial ecosystems. Implemented in MATLAB R2024a and TensorFlow 2.17, the model achieves rapid convergence by epoch 142 and 98 % classification accuracy (AUC = 0.846, MSE = 0.79, recall = 90.6 %) on selected high-liquidity assets. These results confirm the framework’s ability to model nonlinear dependencies and adapt to stochastic disturbances typical of service-oriented and engineering-operation contexts. Beyond predictive precision, the EDNN–LR framework provides explainable probabilistic outputs that can be directly incorporated into decision variables such as resource allocation, demand forecasting, and dynamic scheduling under real-time constraints. Its hybrid design reduces computational cost, enhances interpretability, and enables cross-domain adaptability—from financial risk management to logistics, supply-chain coordination, and energy-system optimisation. By bridging deep learning, system engineering, and behavioural decision analytics, this study contributes a generalised AI-driven architecture for intelligent and transparent decision-support across Industry 4.0 service and production ecosystems.