The distributed simulation of highly automated batch manufacturing systems
This thesis examines the use of distributed discrete-event simulation techniques as part of an aid to the design of highly automated batch manufacturing systems. The methodology and objectives of the design of highly automated batch manufacturing systems are described and an assessment is made of the use of modelling and simulation as part of the method . Criteria are developed for a simulator used during the . detail design stage. The different approaches taken by existing simulation systems to building and configuring simulation models and their use of particular simulation techniques are described. Limitations on simulation models due to the sequential processing of event-lists and activity scans are identified in a review of the problems of simulation that current existing distributed simulators have been designed to answer. The advantages of concurrent and distributed computing and in particular, a tightly-coupled multi-microprocessor computing engine for executing the normally batch-processed computing tasks of simulation are identified . A novel approach to the distr~bution of the computational tasks in a distributed simulation system is described and the operation of a simulator built using this approach to simulate the operation of highly automated batch manufacturing systems is also described.The question of whether such a distributed simulator of highly automated batch manufacturing systems satisfies the criteria is examined on the basis of an analysis of the operation of the simulator. It is shown that a number of advantages in the areas of level of detail, configuration, parallel processing and speed of execution can be achieved through the use of distributed computing and multi-processing techniques for simulation during the detail design stage of highly automated batch manufacturing systems.