Choosing the appropriate capacity for a make-to-forecast production environment using a Markov analysis approach

In the "make-to-forecast" production environment, competitive market dynamics require customer delivery times substantially shorter than the fixed manufacturing lead times, which require the release of units into production without prior knowledge of customers' desires. As a result, there is the possibility of either getting more orders than can be accommodated causing the rejection of some, or getting too few orders leading to a finished unit without a buyer, which we term an "orphan". The physical size and dollar value of the units make storing of the orphans, if not completely impossible in some situations, at least extremely undesirable. The likelihood of these two undesirable events depends on the managerially predetermined production capacity relative to the exogenous average order arrival rate. If the capacity is excessive, too many units will be orphaned, whereas insufficient capacity will result in too many rejected orders. We present a Markov model to analyze the behavior of the system in regard to orphan and order rejection levels. The analysis provides management and researchers with highly generalizable insights into managing this common dilemma under various demand and policy scenarios to make more informed capacity level decisions.