Nonstochastic Analysis of Manufacturing Systems Using Timed-Event Graphs

Using automated methods to analyze the temporal behavior ofmanufacturing systems has proven to be essential and quite beneficial.Popular methodologies include Queueing networks, Markov chains,simulation techniques, and discrete event systems (such as Petrinets). These methodologies are primarily stochastic. Performanceevaluation mandates results which are probabilistic in nature (such asthe average rate of part deliveries) and relies on probabilisticinputs (such as the probability of breakdown, or the distributionsassociated with a manufacturing process). This paper examinesnon-stochastic analysis, which we argue can be useful for verifying{\em correct} operation. We model manufacturing systems using timedevent graphs which are similar to decision free Petri nets augmentedwith timing information, and present an example that demonstratesthe efficacy of non-stochastic analysis.