Identification of potential deadlock set in semiconductor track systems
Hyun Joong Yoon, Doo Yong Lee · 2002
Addresses an efficient method to identify the process modules that can potentially cause deadlock in semiconductor track systems. Track systems carry out important functions in the photolithography process of semiconductor fabrication. Trends toward high automation and flexibility in the track systems accelerate the necessity of the intelligent controller that can guarantee deadlock-free operation of the track systems. However, there is a limitation when applying a deadlock avoidance policy in real-time due to the computational complexity caused by scores of modules, and complex and flexible process flows. To cope with that problem, the paper proposes an efficient method to identify potential deadlock process modules by introducing the concept of a potential deadlock set. The potential deadlock set is defined as the set of resources or modules that can result in deadlock in near future. Once the potential deadlock set is identified in the track system, a deadlock avoidance policy can be applied only to the process modules included in the potential deadlock set. Therefore, the approach using the potential deadlock set can significantly reduce the computational complexity when applying deadlock avoidance policy. A resource request matrix and a genetic algorithm are used to find the potential deadlock set in the track systems.