Performance evaluation of an auction based manufacturing system using generalized stochastic Petri nets

Auction Based Manufacturing is a form of Heterarchical control. Such a system relies upon the intelligence of individual machining centres to allocate parts to prospective machining centres which can do the next stage of machining. This approach eliminates the requirement of a costly pre process scheduling run, which can prove to be inappropriate in the event of unforseen machine breakdowns or changes in machining capacities.;One of the major drawbacks is the lack of predictability of the exact behaviour of the shop floor. It is perceived that a method by which an operator can predict the most likely behaviour of the system is highly desirable in order to fine tune the control algorithms or rules to achieve optimality and make the right resource available at the right time for the right amount of duration.;This thesis deals with the modelling and the performance evaluation of an Auction Based Manufacturing system that has multiple machining stages, multiple part types and multiple parts. Each part type has a process plan with multiple stages, further each stage can have multiple alternatives available. Four different models have been analyzed to understand the impact of control heuristics, tie-breaking rules, number of parts and machining rates on the overall system behaviour. (Abstract shortened by UMI.).