Bullwhip Effect in Supply Networks : The Impact of Network Structure

The bullwhip effect refers to the amplifcation of demand variation when moving upstream the supply chain. Bullwhip research so far has mainly focused on linear and distributor networks while integrator and more general network types have not been analyzed in detail yet. First, we derive a metric that allows us to measure the bullwhip effect in a general network structure. Second, we quantify the bullwhip effect in different network archetypes (linear, diverging, integrating and generalized networks). Third, we simulate the bullwhip effect both for network archetypes and for a large number of real supply networks with distributor, integrator and generalized structures.The bullwhip effect can be shown to decrease strongly if network settings are heterogeneous (i.e., nodes use different number of forecasting periods and order scheduling) and if horizontal complexity (i.e., number of suppliers per network tier) increases. In real supply networks, the bullwhip effect tends to be much lower (up to -97% for selected structures) than in linear supply chains. For network types that are vulnerable to the bullwhip effect, we propose strategies to reduce it