Cycle time learning curve in semiconductor foundry industry

Faster cycle time is considered to be one of the key factors for the competitive semiconductor foundry industry. During the past decades, many studies had tried to provide rules and models to guide activities for reduction of cycle time. [Yu Hsin Lin et al.] [Wallace J. Hopp et al.] Most of studies aimed on Issue of WIP/tool/dispatch and line balance to reach a cycle time target. But, what's a good cycle time target? How to set a good cycle time target? Learning curve provide us a good way to answer these questions and show us how powerful it is. In this paper, we establish cycle time road map by applying learning curve approach [Benjamin W. Niebel] to find an empirical rule. Based on this rule, we can define cycle time goal, road map, and long-term activities to make the most profits for the company. We collected historic data and used regression function to find the learning rate. Using the rate we set a tough but reasonable target to drive C/T reduction activities to have a most competitive cycle time