Analytical Performance Analysis of Kanban Systems Using a New Approximation for Fork/Join Stations

In this paper, we propose a new approximate approach for analyzing queuing models of single stage kanban systems. The approach is based on parametric decomposition technique and two moment approximations. The main building block of our approach is a new two-moment approximation characterizing the performance of fork/join stations fed by general inputs. Using this approximation we develop analytical methods to estimate different measures of kanban system performance such as throughput and work in process.