Intelligent Value Stream Mapping with Embedded Predictive Analytics

Abstract

This review examines Intelligent Value Stream Mapping (VSM) with embedded predictive analytics as an advanced lean manufacturing approach for overcoming the static, retrospective, and manually dependent limitations of conventional VSM. It focuses on how real-time data, machine learning, digital twins, and Industry 4.0 technologies can transform VSM into a dynamic decision-support system for proactive waste reduction and process optimization. The study synthesizes literature on traditional VSM, predictive analytics, data-driven manufacturing architectures, digital value stream twins, process mining, edge/cloud analytics, and lean tool integration. It evaluates data acquisition from sensors, IoT/RFID devices, ERP, MES, and shop-floor systems, together with preprocessing, feature engineering, predictive modelling, simulation, dashboard visualization, and decision-support mechanisms. The review shows that Intelligent VSM strengthens bottleneck prediction, inventory and lead-time management, continuous kaizen, demand-driven scheduling, and integration with Kanban, Andon, Poka-Yoke, Heijunka, and other lean tools. Evidence from the literature shows measurable gains, including 28% reduction in expected delivery time, 67.84% lead-time reduction, 20% productivity improvement, 25% OEE improvement, production balance improvement of 29.07%, and ANN prediction performance with MSE below 0.001 in selected applications. However, adoption remains constrained by poor data quality, heterogeneous system integration, black-box model behaviour, scalability limits, latency constraints, weak alignment with lean simplicity, and lack of standardized implementation frameworks. Future research should prioritize AI-driven optimization, digital twin ecosystems, edge-enabled real-time analytics, interpretable human-centered dashboards, and scalable standardized methodologies.