Modeling and Simulation of Manufacturing Systems Under Signal Kanban Policies

Lawrence E. Holloway, University of Kentucky

Presented at the 2nd International Symposium on Scale Modeling, Lexington, Kentucky, May 1997.

Abstract

A signal kanban system is used to order the production of various parts by processes with lengthy setup times, such as in metal stamping or injection molding. In this paper, we explore the behavior of signal kanban systems using simulations of systems producing two part types. Such simple models allows a visualization of the changes in the inventories of these parts under the signal kanban system. We consider the operation of the common fixed-batch signal kanban system, and a variation that we refer to as a fixed-fill policy. These policies are investigated under variations in demand, occasional disruptions in production, and modifications in policy parameters. Our investigations for this two-part­production model indicate that a fixed-fill signal kanban system is significantly more robust than the more common fixed-batch kanban system.