Optimal procurement strategies by reverse auctions with stochastic demand

This paper studies a dynamic procurement problem by reverse auction for a retailer with stochastic demand. In each period, the retailer based on his inventory needs to determine a payment function (a procurement contract) according to which a number of potential suppliers compete in the reverse auction. We show the existence of the retailer's optimal payment function and find that the suppliers' Bayesian-Nash equilibrium bidding strategy is similar to the base-stock policy in the traditional multi-period inventory control problems when the retailer incurs no fixed setup cost, while similar to the (s, S) policy when the retailer incurs a fixed setup cost. This strategy is for the suppliers, instead of for the retailer, depends on the supplier's marginal cost and so is stochastic for the retailer. Thus, this paper extends well beyond traditional procurement environments studied so far in the inventory control literature.