2.1 Introduction

Analysis, planning and control of a supply chain calls for a combination of spreadsheet, optimization and simulation models. Spreadsheet analysis is by far the most popular form of supply chain modeling due to its accessibility, ease of use and flexibility. However, spreadsheets are fairly limited in modeling power, with a few notable exceptions (Katircioglu et al. 2002). Optimization technology such as linear or mixed integer programming is a great way to solve well-defined mathematical problems such as supply network planning and inventory optimization (Ettl et al. 2000). But optimization models are rigidly structured and often based on simplifying assumptions to make the problem fit the mathematical format required by the underlying solver. Another issue that often limits the utility of optimization is uncertainty. Uncertainty abounds in supply chains for example in customer demand, lead times and supply availability. Although optimization under uncertainty is a popular research topic, few commercial supply chain optimization tools support uncertainty models.

Simulation is a popular alternative to optimization for supply chain analysis. Simulation models are not restricted by rigid mathematical structures. Almost any supply chain issue can be coded as a simulation object with a set of parameterized behaviors. Individual components of a supply chain can be modeled separately and then combined into one large simulation model to represent the overall system. With simulation it is relatively easy to incorporate uncertainty, by generating random numbers for uncertain parameters during simulation runs. Multiple iterations are required to understand...