DEVELOPMENT OF POLICIES FOR DEADLOCK DETECTION AND RESOLUTION, AND AVOIDANCE IN AUTOMATED MANUFACTURING SYSTEMS

Abstract

A deadlock is a situation where each of a set of two or more parts keeps waiting indefinitely for the other parts in the set to release resources. Deadlocks may arise as the final state of a ,complex sequence of operations on concurrent parts passing through a system, and are thus generally difficult to predict. Deadlocks lead to degraded performance and ultimately zero throughput. The automated operation of an AMS makes deadlocks an important problem to be tackled. In this work, first stage focuses on the development of a discrete event simulator for an automated manufacturing system. In the second stage, the basic simulator was tailored to detect and resolve deadlocks. In the third stage, again the basic simulator was tailored to avoid deadlocks. There are two cases in each of these methodologies. In the first case, manufacturing system is capable to handle only one fixed route for each part type, while in the second case, it can handle two alternate routes. Software is developed in C language and is user friendly. Several case studies are taken to demonstrate the effectiveness of the proposed with both single and multiple routings. In addition, the effect of buffer capacities is also studied under these policies. Simulation results indicate that utilization of machines increases with the increase in buffer capacity and alternate routings give better system performance than single routing. Depending on the suitability one of these deadlock handling policies can be selected for the manufacturing system.