Journal Article

Dynamic Multisourcing with Dependent Supplies

Feng, Qi, Jia, Justin, Shanthikumar, J. George

2019

Overview

As multisourcing becomes a widely adopted strategy in dynamic procurement planning, firms inevitably source from suppliers with dependent material flows, and such supply dependence, created by common second-tier suppliers or a common economic environment, is often positive. We show that firms can experience significant profit loss when implementing a policy computed based on a model that ignores the dependence among suppliers’ stochastic material flows. The profit loss is particularly significant when a firm earns a low margin or faces an unreliable and fragmented supplier base. Analysis of the firm’s dynamic multisourcing strategy under dependent supply uncertainties, however, is challenging. By expanding the space of ordering decision to a class of contingency policies, we identify conditions under which the dynamic planning model has a concave transformation. This approach allows us to characterize the optimal procurement policy. Moreover, our analysis suggests that the firm may order more from a less reliable supplier than from a more reliable one who charges the same procurement cost. The order allocation to the less reliable supplier can be even larger when the capacities are more dependent. Though the analysis of our base model is conducted for dependent supply capacities that are independent across periods, we show that the results can be generalized to Markov-modulated supplies or general supply functions that are stochastically linear in midpoint. This paper was accepted by Noah Gans, stochastic models and simulation.

• Publisher: INFORMS,Institute for Operations Research and the Management Sciences
• Subject: contingency policies / Markov analysis / multisourcing / positive dependence / Purchasing / sequential stochastic monotonicity / stochastic linearity / Stochastic models / stochastic supply functions / Suppliers / Supply chains / Transformation