Kanban, MRP, and Shaping the Manufacturing Environment

Abstract

There has been a great deal of interest recently in the Japanese approach to manufacturing, growing out of a concern for finding ways to reduce inventories and increase productivity. At this project's inception, its objective was to assess whether the kanban system could perform well in the manufacturing environments found in this country. Based on observations from managers visiting Japan, the project was enlarged to also assess which factors in a production environment have the biggest impact on performance—regardless of the system in use. Guided by a panel of production and inventory managers from diverse plant environments, a comprehensive list of factors thought most important to manufacturing effectiveness was constructed. The panel established low and high values for each one. These settings were considered representative of the range experienced in U.S. plant environments. The factor settings allowed a variety of representative plants to be tested with a large scale simulator. Results show that kanban, when implemented in certain environmental settings, does indeed perform exceptionally well. However, so do the more traditional systems used in the United States. Conversely, there are other plant environments in which all systems perform much worse. This suggests that the factors themselves are the keys to major improvement. Simultaneously reducing setup times and lot sizes is found to be the single most effective way to cut inventory levels and improve customer service. Shop factors of particular importance are yield rates and worker flexibility. Degree of product standardization and the product structure are also high impact factors. Less crucial than earlier believed, at least over the factor settings simulated, are inventory record inaccuracy, equipment failures, and vendor reliability. Such results suggest that the selection of a production/inventory system can be of less importance than the improvement of the manufacturing environment itself.

Cite this work

No public copy is stored for this record, so only the metadata is cited.

Stored only when both papers are in the corpus — 387 citations exist upstream.

Papers that cite the same references — the relational fallback while embeddings are still backfilling.