Comparison of approaches to reorder level calculation in a continuous replenishment strategy

Based on a first order autoregressive and moving average (i.e. ARMA(1,1)) demand process, an optimization model of the variable reorder level in a continuous replenishment strategy is developed. A theoretical comparison is conducted to demonstrate that when the autocorrelation coefficient and moving average coefficient are positive, the safety stock with a variable reorder level is less than that determined by means of moments of demand, but more than that with traditional reorder point approach. Furthermore, the safety stock level difference between using a variable reorder level based on moments of demand, as well as that between using a variable reorder level based on traditional reorder points are increasing in replenishment lead time, demand autocorrelation degree and random error moving average coefficient. Finally, a numerical simulation experiment is made to compare the inventory costs with different models. It is shown that the average stockout rate and average inventory cost based on variable reorder level are less than those based on the other two reorder levels.