Description of the Computer Program for METRIC — A Multi-Echelon Technique for Recoverable Item Control
A user's manual including program printout, for the METRIC system of optimizing Air Force depot and base stock levels of high-cost low-demand items. Since demand for recoverable items is randomly distributed, METRIC assumes a logarithmic Poisson distribution of demand for each item and applies Bayesian inference to assign probabilities to the set of possible future demands. The program can handle a system of one depot and up to 20 bases, assuming that base inventory policy is to reorder one unit whenever one is demanded; the criterion is minimal expected backorders at base level, considering backorders in terms of days of unfilled demand for any item at any base. The program ignores lateral resupply between bases, does not allow batching of items for repair or resupply, and assumes all items to be reparable either at base or depot level. New procurement is optimized, or existing assets are redistributed optimally, so as to (1) minimize backorders for a given investment, (2) minimize investment for a given backorder level, or (3) apply a predetermined Lagrange multiplier to the ratio of decrease in backorders/unit cost. Fill rates are also computed.