A lean smart framework for predictive agility in manufacturing using action research

The integration of Smart Technology with Lean Manufacturing enhances operational sustainability. However, when digital technologies are incorporated into unstable physical processes, they often generate 'digital waste' rather than actionable intelligence. This research addresses the gap between physical and digital technologies by proposing and validating a sequential 'three tier architecture' approach that establishes lean physical stability as a strict prerequisite for cyber-physical connectivity and mathematical optimisation. This research utilised Action Research methodology, the study sequentially implemented flow-control mechanisms (5S/Kanban), the Internet of Things for condition monitoring of manufacturing processes, and Mixed-Integer Linear Programming (MILP) to execute a block allocation decision model using Python, sequentially. Based on empirical evidence from the implementation, the integrated approach outperformed siloed implementations. The predictive model reduced the wear-related waste from 28.67 to 9.07%. The MILP optimisation increased block utilisation by 10.72% and reduced total process waste from 15.28 to 10.35%. This research extends the theoretical paradigm of 'Lean 4.0' beyond existing definitions by proposing that physical stability moderates 'Predictive Agility' in manufacturing environments. Furthermore, it provides small and medium-sized enterprises with a cost-effective framework to increase material circularity and improve operational reliability without investing in extensive, techno-centric solutions.