Research Papers on Operations & Supply Chain Management

We present a new approach for decentralized coordination in light of asymmetric information. The basic idea is to perform an iterative exchange of primal information, like proposals for purchase and supply quantities, in order to identify improvements over an initial, uncoordinated solution. Unlike most existing approaches, it will coordinate decisions of two or more decentralized parties, provided that the decision problems can be modeled and solved as linear or mixed-integer programs, while it does not require a central decision-making entity. We present a coordination scheme that identifies the system-wide optimum in a finite number of iterations, provided that all but one decentralized party report their cost changes incurred by new proposals and that the models of these parties can be formulated as linear programs, while the model of the remaining party may be of a mixed-integer type. In computational tests, we find that 20 iterations are sufficient to obtain considerable improvements over initial master plans in supply chains consisting of one buyer and one or multiple suppliers. Moreover, we show that the information exchange can be embedded into a strategy-proof mechanism, for which a lower bound on the efficiency can be derived.