Material flow control in Remanufacturing Systems with random failures and variable processing times

Abstract Material flow control in remanufacturing is an important issue the field of disassembly. This paper deals with potential autonomous material release decisions for systems to balance uncertainties related changing bottlenecks, maximise throughput ( $$TH$$ TH ) and minimise work-in-process $$WIP$$ WIP ). The goal achieve highest possible rate using real-time data while keeping costs a minimum. Unlike traditional production systems, must consider handle high process. Up now, classical methods such as CONWIP, Requirement Planning (MRP) Kanban have been used control. However, these do not perform well system variation they aim find solutions static environments. Crucial optimal stochastic environments finding optimum pull or which can vary over time terms maximising minimising . We propose deep reinforcement learning approach that acts on environment adapt conditions. ensures bottlenecks are taken care there minimum system.