Analytical approximations for kitting systems with multiple inputs

This research analyzes a kitting system where multiple components are grouped to form a predefined kit prior to assembly. The kitting system is modeled as a fork/join synchronization station and component supply is assumed to be from fabrication facilities operating under a kanban control policy. Exact analysis of these systems is computationally intensive even under Markovian assumptions. To evaluate the impact of input parameters on kitting system performance, several easily computable bounds on system throughput are first identified. Using these bounds, closed form approximate expressions for system throughput are derived. The throughput estimate is used to compute other performance measures of interest such as mean queue length and mean waiting time in the system. The accuracy of the approximations is validated using numerical experiments and some performance insights are given.