Enhancing supply chain efficiency in textiles: a deep learning approach to Industry 4.0 implementation

Moroccan textile SMEs face increasing pressure to adopt Industry 4.0 (I4.0) technologies to enhance their competitiveness and improve their supply chains. However, a lack of clear implementation strategies, particularly regarding action prioritization, hinders their progress. This research addresses this challenge by developing an intelligent framework, grounded in Deep Learning, to guide I4.0 implementation for these SMEs. The framework leverages two key inputs: the Smart Industry Readiness Index (SIRI) dimensions, providing a structured assessment of the enterprises’ current maturity across Process, Technology, and Organization, and Critical Success Factors (CSFs), identified through the DEMATEL method, capturing expert knowledge on the drivers of successful I4.0 adoption. The core of the framework is a set of specialized neural network architectures, trained to forecast the appropriate priority domain for I4.0 deployment. These specialized models, including the RNN with Attention (for Organization Priority) and the CNN-LSTM with Attention (for Technology and Process Priority), enable a nuanced and context-aware prioritization of actions. Key performance indicators show high accuracy in determining the appropriate focus area for Industry 4.0 deployment: the RNN with Attention architecture achieves 84.3% accuracy for Organization Priority, the CNN-LSTM with Attention achieves 93.3% for Technology Priority, and 83.3% for Process Priority. This data-driven and expert-informed prioritization approach offers a practical and actionable roadmap for Moroccan textile SMEs to optimize their limited resources and maximize the impact of their digital transformation efforts, ultimately contributing to a more competitive and sustainable sector.