Comparing kanban control with the theory of constraints using Markov chains

Two kinds of drum–buffer–rope (DBR) system, which differ according to the place to which orders are directed, are proposed. A queuing-network model of a kanban control system (KCS) and two kinds of DBR system for a three-stage serial-production system with a bottleneck were developed. The performances of the KCS and the two kinds of DBR systems are compared by using Markov analysis. The average number of stocked and in-process inventories, the average number of stocked products, the average number of rejected demands, and total cost are considered as performance measures. Numerical experiments are conducted to investigate the influence of the processing rate and cost parameter on the performance measures. The results show the advantages of each system over the others.