The impact of market state information on inventory performance

Abstract

This research studies the impact of market state information on the long run average cost of a single item inventory system facing dynamic, random demand. The probability distribution of the demand in any time period is determined by the current market state in that period which is modelled as a Markov chain. The inventory manager must infer the market state from the demand observations and a market state signal that is not necessarily accurate. The proposed model accommodates a range of different levels of accuracy for this market state information. The computational results indicate that the value of the market information increases at an increasing rate as the accuracy increases suggesting a strong, persistent value from increasing the information accuracy. The computational results also characterise the effects of lead time, the difference between the market state distributions, autocorrelation, and service level on the value of the market state information.