Analytical performance modeling of supply chain networks

Scope and method of study. The purpose of this study was to develop queueing-based analytical models of building-block configurations of supply chain networks based on the parametric decomposition approach, and to use these models to develop the model of a more complex supply chain network. Existing queueing models of production-inventory networks were extended by including supply operations, transportation and distribution. Different building-block configurations such as a divergent network, a convergent network with two manufacturers, and a convergent network with two suppliers were modeled. In modeling a convergent-divergent configuration, which was a combination of the building-block configurations, the analytical expressions from the models of individual configurations were used. Validation of the analytical models was done by comparing analytical results to simulation estimates obtained using Arena simulation models.; Findings and conclusions. The analytical models were found to yield reasonably accurate results and the relative percentage errors were well within the permissible error of 20% for a majority of the cases tested. This research also showed that the building-block approach works well when applied to larger and more complicated networks. This work has laid the foundation for the development of rapid analysis tools for supply chain networks, and has successfully extended the popular parametric decomposition approach based on two-moment approximations to model supply chain networks.