In multiproduct manufacturing systems, it is difficult to assure that an optimised setting of a pull production control strategy will be able to maintain its service level and inventory control performances. This is because the competition for resources among products is liable to make them affect the service levels of one another.
By comparing different pull strategies, this research has observed that tightly coupled strategies are able to maintain lower amount of inventory than decoupled strategies, but they do so at the detriment of service level robustness. As a result, tightly coupled strategies are better suited to manufacturing environments with low variability, while decoupled strategies are more robust in high variability environments. Here, robustness is a measure of how well a strategy is able to minimise the drop below its original optimised service level when the initial system conditions change.
Furthermore, the Kanban allocation policy applied under a strategy plays a major role in its ability to manage the performances of multiple products. Experimental results show that the Shared Kanban Allocation Policy (SKAP) keeps a lower amount of inventory than the Dedicated Kanban Allocation Policy (DKAP), but it is more susceptible to the variability in the demand or processing times of one product impacting the service level of another. Therefore, a Hybrid Kanban allocation policy (HKAP) that combines both the DKAP and the SKAP has been implemented. This approach considers products’ demand and processing time attributes before categorising them into the same Kanban sharing group. The results of the implementation of the HKAP show that it can keep as low inventory as the SKAP and avoid products impacting the service levels of one another. Additionally, it offers a better approach to managing large multiproduct systems, as the performances of product groups can be differentially managed through the combination of Kanban sharing and dedication policies.
Lastly, the observations on the performances of strategies and policies under different system conditions can be used as a framework through which line designers select strategies and policies to suit their manufacturing system.