| Resources that have a restriction on availability can lead to deadlock conditions in an automated manufacturing system (AMS). In a flexible manufacturing system (FMS) environments, machines, buffers, materials handling systems, and other auxiliary items such as tools, robots, fixtures, and pallets can be included in the restricted resource category. In general, deadlocks are difficult to predict. The earlier a deadlock or its possibility in the system is detected, the easier it can be resolved or prevented, and consequently, the smaller is the lost production. Whether the deadlock problems can be efficiently resolved at one level or the other is not entirely clear and is somewhat philosophical. In this paper, three different research papers are presented. Each paper deals with deadlock problems but takes different approaches at different levels to resolve deadlocks in FMSs.;In the first paper, a real-time control algorithm is proposed for deadlock resolutions in more general AMSs. This algorithm is based on a graph theory and "module" concept which is also defined and discussed in this paper. A module refers to a set of resources that can be replaced as a whole in deadlock detection. By grouping resources before real-time operation, the complexity and time to execute the detection of deadlock in real-time can be minimized.;In the second paper, two real-time deadlock resolution models are proposed for FMSs with automated guided vehicle system (AGVS). The two models use the characteristics of the structure of FMSs with an AGVS. The two models proposed are based on the type of AGV dispatching policy employed in the normal operation.;In the third paper, a buffer management model for an FMS with an AGVS is proposed and tested. Simulation experiment results clearly indicate that the proposed buffer management model is superior over the conventional buffer management model in terms of throughput and the required capacity of the buffers to prevent deadlocks in the systems. |
