Estimating average production intervals using inventory measurements.

This memorandum proposes an indirect approach for estimating average production intervals using work-in-process inventory measurements. The idea is to apply a modified version of Little's Law (L=lambda W) from queuing theory to cope with stochastic processes that are not directly observable. When the actual amount of completed product that will eventually be produced from current work in process is not known, we suggest working with an appropriate expected amount of completed product associated with current work in process, taking care to properly account for such features as partial yields, changing lotsizes and reconstitued lots. This indirect estimation procedure can be applied with computer simulation as direct system measurement. The approach can also be used to calculate expected. values of steady-state random variables.