Comparing JIT approaches in a manufacturing cell

Traditionally, production control in Just-in Time (JIT) systems is designed to replace an item that has been consumed at a downstream work center by 'pulling' a new item through the manufacturing process to that center. However, when information on future customer orders or production requirements is available, there is no reason the system can't pull an item through the process to satisfy a future need. This article compares these two approaches to the design of a kanban-controlled pull system for a manufacturing cell. The cell is patterned after one that produces engine blocks that feed directly onto the assembly line at a European plant manufacturing large diesel engines. A simulation model is used to evaluate the cycle times, capacity utilization, and output rates for the two approaches. The comparison is conducted for a number of inventory, uncertainty, and protective capacity levels in the cell. The results show that a JIT system that anticipates future needs is superior to one that replaces items as they are used. In addition, the results suggest that, in either approach, variance reduction and capacity increases in the non-bottleneck stations are serious alternatives to using inventory for buffering.