Applied Research On Factor Screening Methods In The Simulation Of Supply Chain

It is important to identify important factors in simulation experiments. Many controllable or uncontrollable factors are included. It requires a lot of simulation time and much cost if all factors are tested. It simplifies the process in the next step of the simulation to reduce the time and the cost, if important factors are found out by screening all factors in the initial simulation. And the researchers can prepare for the robust design for the supply chain. Therefore, it is more and more important to research on the screening methods.This thesis describes the background and the significance of researching on the screening factors in the simulation, and summarizes different screening methods of domestic and foreign scholars. Then it briefly introduces the basic theory and concept related. It introduces and analyzes the sequential bifurcation screening methods in the third chapter. In addition, it describes a supply chain simulation model of the multi-echelon inventory system. The important factors are screened by sequential bifurcation screening method in this model. The sequential bifurcation method is improved in the next chapter and proved to be correct and effective. The new method is used in the same simulation and identifies the list of the important factors. It is more efficient by comparing the screening results of the two methods. The improved method is used in the simulation model that has multiple output at the end of the thesis.The innovative work in this dissertation is summarized as follows:1) Factor screening method is rarely applied in the simulation of a supply chain. The research riches the theoretical research of the factor screening method in the supply chain simulation.2) The sequential bifurcation and the improved method were applied to a food supply chain simulation of the multi-echelon inventory systems.3) The sequential bifurcation method was improved to identify the important factor. It makes up the shortage of the initial method and expands the theoretical research achievements.4) It was initially discussed to screen factors in the simulation model which have multiple output. It can expand the application range of the screening methods.