Parameter Estimation for Two-machine Non-exponential Production Line Models using Parts Flow Data

The foundation of any quantitative analysis and optimization for a production system is a high-fidelity mathematical model of the physical process. To construct a stochastic process model more efficiently, a new approach to production system modeling has recently been proposed that uses system performance metrics (e.g., throughput, work-in-process) derived from the parts flow data to reversely compute the model parameters. This paper extends this modeling approach to two-machine non-exponential production lines. Since no analytical expressions of performance metrics are available for non-exponential systems, we use neural network-based surrogate models as a means to calculate those performance metrics as functions in terms of the system parameters. Then, based on the constructed surrogate models and given performance metrics, the machine parameters are estimated by solving a constrained optimization problem using a multi-start particle swarm optimization algorithm. The results indicate that different combinations of machine parameters can lead to practically the same system performance metrics and a linear relationship of the reliability parameters from these obtained estimations is observed. Numerical experiments are implemented to justify the robustness of the different combinations of machine parameters.