A multicriteria framework for inventory classification and control with application to intermittent demand

Abstract

Abstract Several papers have studied inventory classification in order to group items with a view to facilitating their management. The generated classes are then coupled with the specific reorder policies composing the overall inventory control system. However, the effectiveness of inventory classification and control system is strictly interrelated. That is to say, different classification approaches could show different performance if applied to a different set of reorder policies, and vice versa. Furthermore, when the cost structure is subjected to uncertainty, a pure cost‐based analysis of the inventory control system could be corrupted. This paper presents a multicriteria framework for the concurrent selection of the item classification approach and the inventory control system through a discrete‐event simulation approach. The key performance indicators provided by the simulator (i.e., average holding value, average number of backorders, and average number of emitted orders) are indicative of the multidimensional effectiveness of the adopted inventory control system when coupled with a specific classification approach. By this way, a multicriteria problem arises, where the alternatives are given by exhaustively coupling the item classes, which are generated by different classification approaches, with the reorder policies composing the inventory system. An analytical hierarchy process is then used for selecting the best alternative, as well as for evaluating the effect of the weights assigned to the key performance indicators through a sensitivity analysis. This approach has been validated in a real case study with a company operating in the field of electrical resistor manufacturing, with a view of facilitating the management of items showing intermittent demand.

Cite this work

No public copy is stored for this record, so only the metadata is cited.

Stored only when both papers are in the corpus — 18 citations exist upstream.

Papers that cite the same references — the relational fallback while embeddings are still backfilling.