Towards online reinforced learning of assembly sequence planning with interactive guidance systems for industry 4.0 adaptive manufacturing

Literature shows that reinforcement learning (RL) and the well-known optimization algorithms derived from it have been applied to assembly sequence planning (ASP); however, way this is done, as an offline process, ends up generating methods are not exploiting full potential of RL. Today’s lines need be adaptive changes, resilient errors attentive operators’ skills needs. If all these aspects evolve towards a new paradigm, called Industry 4.0, RL ASP needs change well: phase has part execution optimized with time several repetitions process. This article presents agile exploratory experiment in prove effectiveness techniques execute adaptive, online experience-driven directly at time. The human-assembly interaction modelled through input-outputs guidance system built digital twin. Experimental assemblies executed without pre-established plans adapted experiments show precedence transition matrices for can generated statistical knowledge different executions. When frequency given subassembly reinforces its importance, results obtained applications only possible but also effective learning, teaching, executing improving tasks same paves application ASP.