Simulation Models for Decision Making

Uncertainty leads to challenges in decision making. Mathematically, we represent uncertainty by defining probabilities when several of the outcomes are possible in the future. This modules provides an overview of probability concepts that are essential to lay a good foundation for simulation modeling. We will also get our first exposure to Excel based simulations.

While being able to estimate probabilities using mathematical relationships is important, a lot of natural events follow or approximate some nicely defined probability distribution functions such as Uniform, Exponential and Normal Distributions. To effectively build simulation models, it is important to understand how to use these distributions. Further, we may need to find what distribution does our observed data follow. This module introduces the finer details of working with probability distribution functions and introduces the types of simulation models as well as some practice based tricks to work with real-world data that may not be complete or may not fit a given distribution exactly.

We started by stating that simulation is one of the most flexible modeling approaches. This module demonstrates that flexibility. In this module, four Monte Carlo simulation models are built for a coffee shop. The models increase in technical complexity and sophistication to demonstrate various issues that modelers have to consider in building these models depending upon the type of questions that need to be answered. The lessons explain which models can answer certain type of questions and what questions may not be answered by a certain type of model. The results obtained from various models are then compared and discussed to understand the tradeoffs in choice of a particular model choice.