AN TERATIVE REFINEMENT APPROACH TO FITTING CLEARING FUNCTIONS TO DATA FROM SIMULATION MODELS OF PRODUCTION SYSTEMS

We examine the problem of fitting clearing functions that estimate the expected output of a production resource as a function of its expected workload from empirical data. Unlike most regression problems, the independent variables are not directly controllable due to the presence of a planning model that controls releases to the production system, and the release decisions made by the planning model are themselves dependent on the estimated clearing function. We propose an iterative refinement procedure that uses simulation experiments to resample data from the production system as the parameters of the clearing function are iteratively updated. We compare the iterative procedure to previously used approaches with promising results.