How the law of unanticipated consequences can nullify the theory of constraints: The case for balanced capacity in a semiconductor manufacturing line

The Theory of Constraints (TOC), described by Goldratt in The Goal, was developed and used to help maximize manufacturing asset utilization. This theory proposes that by intentionally creating a constraint or bottleneck in the line at a particular toolset, several advantages can be achieved. First, by creating such a constraint fewer tools must be purchased. Thus, if the constraint is created on the most expensive toolset, then significant capital investment savings can be realized. Secondly, it is believed that by creating a single constraint or bottleneck in the line, the performance of the entire line will be determined by this toolset and, therefore, from a line management perspective, it is easier to manage a single known constraint rather than multiple constraints. In fact, the belief is that an increased focus on a single bottleneck toolset will generate additional efficiencies for this toolset. This paper explores the consequences of adding a cycle time objective to the performance requirements of a semiconductor manufacturing line. The conditions under which the Theory of Constraints should and should not be used is developed. For most realistic conditions in semiconductor manufacturing lines the TOC will be shown to require a higher asset investment and result in more inappropriate management focus than creating a balanced line with no planned bottlenecks.