Reactive just-in-time ordering system for changes in the mean and variance of demand

This paper proposes a reactive just-in-time (JIT) ordering system for multi-stage production systems with unstable changes in demand, not only in the mean but also in the variance. For a JIT ordering system, that is, the Kanban system, a reactive controller of buffer size is proposed. In the proposed system, unstable changes in the mean and variance of demand can be detected by exponentially weighted moving average charts for mean and variance, and the buffer size at each inventory point is controlled as a reaction to the detected unstable change. The performance of the proposed system is analyzed by simulation experiments.