Discrete Systems Simulation of Production-line Management Techniques: Traditional, Jit, and Toc

This study built a simple, linear, four or fivecell production line model using the SIMAN/ Arena modeling tool. We used information from a case study of a real-world production line, augmented, as necessary, with details of machine operating characteristics not available in the original case study. The model was created in five variants: Traditional, JIT without kanban, JIT with kanban, TOC applied to one workstation only, and TOC applied to two of the workstations. In absolute terms, JIT with kanban produced the greatest improvement in combined plant throughput and WIP -about 13%. This improvement had a cost -the entire production line was restructured and each of the workstations was placed under a SMED/self-inspection/kanban regime. On the other hand, the simplest TOC approach -improve the bottleneck and ignore the rest (for now) -resulted in about a 9% improvement at less than 20% of the cost, cost being measured in terms of number of workstations moved/reconfigured. TOC techniques applied to the two worst bottlenecks resulted in 97% of the improvements expected from JIT, at less than half the cost. INTRODUCTION In the past decade or so two new philosophies of production line management have received great publicity as alternatives to what might be called the Traditional approach: Just-in-Time (JIT), and Theory of Constraints (TOC), also known as Synchronized Production. JIT is a collection of techniques to be applied continuously and factory-wide to reduce Work-in-Process (WIP) and improve production line throughput. With its tight internal linkages (enforced by procedures such as 'kanban' card flow control), and its efforts to reduce setup and downtimes (using techniques like 'Single-Minute-Exchange-of-Die' SMED), JIT tries to make the standard production line look more like a continuous process plant (Chan and Tang, 1994; Chu and Shih, 1992). TOC, on the other hand, tends to look at specific parts of the production process -those which are acting as bottlenecks (Chakravorty and Atwater, 1994; Goldratt, 1990; Leubbe and Finch, 1992; Umble and Srikanth, 1990). TOC production improvement efforts, like JIT, are applied continuously, and ultimately may reach all parts of the plant. The major philosophical difference is that TOC concentrates its efforts, rather than (conceptually) trying to improve everything at once. This work parallels that of Cook (1994), but includes effects of setup and maintenance downtimes, and uses data from an actual production-line. This paper will describe the simulation details of each of the production-line management techniques in turn, and report the salient results. A summary of results is provided in Table 1. We will conclude with a description of how the various approaches fare in two measurements of production-line management effectiveness. MODELING THE PRODUCTION-LINE The production system modeled was a simple, four-station, I-type line. In some cases, an inspection station was added to the manufacturing stations. Details of equipment capacities and downtimes were derived from a case study of an antenna fabrication line described in Schemenner (1993), augmented as necessary with information on processes and results from Shingo (1985). Traditional Approach This approach involved the release of two weeks-worth of materials (five pallet loads) as required by the system. Workstation downtime