Performance study of parallel kanban-base stock for a high-mix multi-stage production system with the entrance of rework

The present study evaluates the performance of parallel kanban-base stock (PKB) system to regulate the production of high-mix items in a multi-stage shared-machine facility with the entrance of rework. The high-runner products are handled using a kanban system and low-runner products using a base stock system. The PKB systems were differentiated in terms of the model driver, loading rule, and rework entrance policy. The performance measures adopted the total output, average work-in-process, flow times and average machine utilisations. Discrete-event simulations were conducted with an analysis constituted of multi-factor ANOVA and response surface methodology. From the analysis, there is a specific R(LR) value that gives desired performance, and at the same time, its dependency is solely on accuracy of regression equation. Between rework policies, predominantly merge rework entrance policy yields more desirable results as observed within performance measures, compared to original rework entrance policy.