Stock control analytics: a data-driven approach to compute the fill rate considering undershoots

Abstract One of the most frequently used inventory policies is order-point, order-up-to-level ( s , S ) system. In this system, continuously reviewed and a replenishment request placed whenever position drops to or below order point, . The variable quantity cycle characterize system by use complex mathematical computations. Different methodological approaches diminish complexity neglecting undershoots, i.e., that point when it reached. paper, we conceptually empirically analyse bias undershoots introduces into estimation fill rate. After that, suggest new methodology developed under data-driven perspective uses state-dependent parameter algorithm correct such bias. As result, propose two methods, one parametric other nonparametric, enhance rate estimate. Both named analytics remove are illustrate practical implications hypothesis on performance design s, This research in lost sales context with simulated stochastic i.i.d. discrete demands as well actual data.