A robust (Q, r) inventory model regarding inaccuracy of point estimation of demand status

Most literatures in design of (Q, r) inventory model assume that the expectation and standard deviation of demand are known. These descriptive statistics are derived by point estimation from historical data. While these point estimates used in the design may not represent the true demand status and sometimes may be far from true values. This inaccuracy of point estimation will lead to severe cost penalties. In this paper interval estimation is applied instead of point estimation. A robust design scheme following minimax framework is conducted to find a satisfying (Q, r). We use relative inefficiency to evaluate performance of a certain (Q, r) in the mini-max process. Relative inefficiency refers to the expected difference between the aggregated cost of (Q, r) and the minimized cost given a precise demand status. Cases are established to validate this scheme.