Dynamic Dual-Period Multi-Criteria Inventory Decision Framework (DPMC-IDF): An Empirical Comparison with ABC and ABC-XYZ

<title>Abstract</title> <p> Traditional inventory classification methods like ABC and ABC-XYZ rank stock-keeping-units (SKUs) based on single-period, parametric metrics - cumulative annual consumption value (ACV) for ABC and the coefficient of variation (CV) for XYZ ^[1] . These methods evaluate data over a static full-year period and depend on a non-robust mean, which implies that a sudden spike in demand or a genuine seasonal increase can end up with the same misleading score ^[2] . This can lead to many items being misclassified in the same manner, and the combined ABC-XYZ matrix tends to amplify rather than fix these issues as it could be expected to ^[3] . In response, this paper introduces the dynamic Dual-Period Multi-Criteria Inventory Decision Framework (DPMC-IDF), which aims to solve these challenges through three key changes: it eliminates the assumption of stationarity by dividing the evaluation period into two sub-periods ^[4] , replaces a parametric mean or variance scoring with a non-parametric, percentile ranking of revenue-share and introduces a signed volatility index (VI) that indicates not just if, but when an item’s demand has shifted. Each SKU receives a three-letter code that reflects its classification across the two sub-periods and its volatility band, which then connects directly to a five-tier decision-support Standard Operating Procedure (Appendix 1). The frameworks will be assessed using quantitative metrics: Misclassification Risk Reduction and its normalized index (MRR/MRRI) after their results are obtained from the application on the simulated three-SKU portfolio. </p>