Glossary

# Sensitivity Analysis (SA)

Tags: Glossary

The study of how the variation (uncertainty) in the output of a mathematical model can be apportioned, qualitatively or quantitatively, to different sources of variation in the input of a model.

## What is Sensitivity Analysis (SA)?

Sensitivity Analysis (SA) is a valuable tool in the field of logistics that helps us understand how changes in input variables can affect the output of a mathematical model. In simpler terms, it allows us to explore the impact of uncertainty or variation in the input of a model on its output.

Imagine you have a mathematical model that predicts the delivery time for a package based on various factors such as distance, transportation mode, and weather conditions. However, these factors are not always known with absolute certainty. There may be variations or uncertainties associated with each of these factors, which can influence the accuracy of the model's predictions.

Sensitivity Analysis helps us identify and quantify the sources of variation in the input variables that contribute to the overall uncertainty in the model's output. By doing so, it enables us to prioritize and focus our efforts on the most influential factors, allowing for better decision-making and risk management.

There are two main approaches to conducting Sensitivity Analysis: qualitative and quantitative. Qualitative SA involves a subjective assessment of the impact of each input variable on the output. It helps us understand the direction and magnitude of the effect but does not provide precise numerical values.

On the other hand, quantitative SA involves assigning numerical values to the input variables and measuring their impact on the output. This approach provides more precise and quantitative insights into the sensitivity of the model. Various techniques can be used for quantitative SA, such as one-factor-at-a-time analysis, factorial design, and Monte Carlo simulation.

One-factor-at-a-time analysis involves varying one input variable at a time while keeping others constant and observing the resulting changes in the output. This approach helps us understand the individual effects of each variable but may overlook potential interactions between variables.

Factorial design, on the other hand, allows us to systematically vary multiple input variables simultaneously. By considering all possible combinations of input values, it provides a more comprehensive understanding of the interactions between variables and their collective impact on the output.

Monte Carlo simulation is a powerful technique that involves randomly sampling input values from probability distributions and running the model multiple times to observe the distribution of the output. This approach accounts for the uncertainty in input variables and provides a probabilistic assessment of the model's output.

Sensitivity Analysis is crucial in logistics because it helps us identify the key drivers of uncertainty and make informed decisions to mitigate risks. By understanding how changes in input variables affect the output, we can optimize our logistics operations, improve efficiency, and enhance customer satisfaction.

In conclusion, Sensitivity Analysis is a valuable tool that allows us to assess the impact of uncertainty in input variables on the output of a mathematical model. It helps us understand the sources of variation and prioritize our efforts to manage risks effectively. Whether through qualitative or quantitative approaches, Sensitivity Analysis empowers logistics professionals to make informed decisions and optimize their operations in an ever-changing and uncertain environment.