Systematic review of data modelling methods for system-level disassembly planning and control

Disassembly is crucial in recovering end-of-life products in a circular economy, e.g. before remanufacturing or recycling. Research addresses the challenges of complexity and uncertainty in disassembly planning and control with data-driven approaches based on data models, such as mathematical programming, machine learning, simulation, and digital twins. However, existing approaches consider data models only within their specific field of disassembly planning and control. Their narrow focus limits the compatibility and applicability of the approaches, which must build upon each other to incrementally refine a disassembly plan in practice. This review provides a comprehensive overview of data modelling methods for disassembly line balancing, disassembly sequence planning, disassembly scheduling, disassembly process planning, and reactive disassembly planning based on 376 publications. A structured catalogue is presented, including 83 data modelling methods based on, e.g. graphs, matrices, and Petri nets. Their application is analysed, showing how methods are used in different research fields. Finally, the scope of the methods is compared, considering product, process, resource, and order modelling. The results support the informed selection and combination of appropriate data modelling methods for data model design across all research fields involved in system-level disassembly planning and control.