Simplification strategies for simulation models of semiconductor facilities

Abstract

Production planning using simulation has been gaining popularity as manufacturing routings become more complicated. However, a detailed simulation model that contains hundreds of machines and thousands of operations is far too large and complex to work with. The thesis in this research is that production planners at plant level do not need to see all the details of the shop floor to develop an effective production plan. Simulation model used for production planning can be simplified to ensure maintainability and manageability of the model. Model simplification is done in two stages, namely aggregation and condensation. Model aggregation reduces the number of routings in the model by using representative flows, whereas model condensation reduces the number of elements in a model by eliminating high throughput rate workstations. Both strategies reduce the complexity of a simulation model without sacrificing its accuracy. Experimental results using real‐life data indicate that the simplified model is a valid representation of the detailed model, and can be used for high level production planning. t‐Tests are used to compare the results of the detailed and simplified simulation model. Finally, rules of thumb are developed to standardize the simplification strategy.