A Survey on Robust Deadlock Control Policies for Automated Manufacturing Systems With Unreliable Resources

摘要

Deadlock is a rather undesirable case in automated manufacturing systems (AMSs). The appearance of deadlock can cause the partial or total stagnation of a system. So far, a large number of deadlock control policies have been developed; nevertheless, the majority are dependent on the assumption that allocated resources cannot break down. In the real world, an AMS consists of a set of concurrent production routes that share and compete for a limited number of resources, such as automated material/component handling devices, buffers, robots, and machines. Resource failures occur unexpectedly. If reasonable control does not exist, a simple resource failure can lead an entire system to stagnation, which can cause enormous economic loss. Therefore, researchers have gradually paid attention to AMSs allowing resource failures in recent years. In this paper, we focus on reviewing and comparing various robust supervisory control policies from the perspective of their structural complexity, behavioral permissiveness, and computational complexity. Some potential future directions are explored. This paper provides a reference source of robust supervisory control of AMSs for researchers and practitioners in this area. Note to Practitioners-In automated manufacturing systems (AMSs), resource failures are common. Their occurrences can lead a system to stagnation, which can cause unnecessary downtime and bring vast economic loss for enterprises. To resolve such stagnation issues, a great number of robust supervisory control policies have been proposed for AMSs with unreliable resources. These policies guarantee that a controlled system can continue to progress without deadlock and blocking states even if some unreliable resources fail to work. By a thorough review of existing robust supervisory control policies for AMSs with unreliable resources, this paper compares and analyzes these policies in terms of their structural complexity, behavioral permissiveness, and computational complexity. The goal of this paper is to provide a reference source in the area to help researchers and practitioners choose a suitable method for solving industrial application problems that are subject to resource failures.