Quantifying the Bullwhip Effect in Networked Structures with Nontrivial Topologies

Abstract

The paper analyzes the formation of the bullwhip effect (BE) in logistic systems organized in a networked structure, as a major threat to upholding stability in the face of non-negligible goods transport delay and uncertainty of demand. As opposed to the majority of earlier works, which considered serial connections, here, more complex -- networked -- topologies are investigated. The popular Order-Up-To (OUT) policy is selected as the method governing the goods flow. Taking into account the fact that in the networked structures the classic indicators do not allow one to properly quantify the BE formation, new measures have been introduced. It occurs that, unlike in the serial configuration, in the networked structures, the OUT policy triggers the BE even in the nominal operating conditions. The BE intensity depends on the type of demand, with the Poisson distribution leading to the largest order-to-demand variance increase. The number of connections and the graph density have been recognized as the principal factors behind the BE formation in networked systems.