Models have a variety of uses in manufacturing as they allow exploration of a system with mitigated risk to the existing system and mitigated financial risk. Both analytical models and discrete event simulation models can help elucidate system behavior, but there can be differences in the results of these two types of models.

The objective of this thesis is to examine the differences between results from analytical models and discrete event simulation models. A series of case studies serves to illustrate why analytical models and discrete-event simulation models differ. The creation of a computer tool called a Learning Historian made it possible to efficiently conduct experiments of discrete-event simulation models.

A flow shop with process drift provides one example of differing analytical and discrete-event simulation models. Even after eliminating errors due to different underlying assumptions, there is a difference between the analytical and simulation model results because of the inherent variability in the simulation model.

A two-stage system that evolves from a push production control to a hybrid system to a pull production control system illustrates additional sources of differences between analytical and discrete event simulation models. The results for the two-stage push model and the hybrid pull-push model from the analytical and simulation models generally agree. Significant errors arise for the two-stage pull model because there is no correct analytical model for the two-stage pull model. The results of the push and pull production control models illustrate the tradeoff between customer cycle time and inventory level.