Reduction of variability in a smart shop floor using discrete event simulation

This paper explores the concept of autonomous decision rules of Industry 4.0. The goal is to model and simulate a production system in which products autonomously choose the resource in which they will be processed based on shortest completion time (shortest queue length). For validation and benchmark, a production system previously presented in the literature is modeled. The performance of the autonomous model is evaluated in comparison to a model with dedicated lines for each product. In the previous work considered, a specific interarrival function and only deterministic parameters for the processing times are used, with no consideration of setup times. After validation, the proposed models are simulated with other parameters (different probability distributions for interarrival and processing times) and are extended to include machine setups, according to different rules. Reductions in total throughput times and their variability were observed for the systems with autonomous products, when compared to those with dedicated lines, which indicates greater responsiveness and stability of the shop floor. An optimization model for scheduling was also proposed and applied to the same system and input data of the simulation of autonomous products, for comparison. The results showed that the proposed decision rule based on the queues' length, used in the simulation, can yield solutions of good quality, close to the optimal solution. This paper contributes to extend and generalize the results of the literature and the discussion about variability seeks to motivate organizations to implement systems with autonomous products.