Uncertainty and Sensitivity Analyses Methods for Agent-Based Mathematical Models: An Introductory Review

TL;DR

This paper reviews three methods for analyzing uncertainty and sensitivity in agent-based biological models, focusing on their origins, implementation, and interpretation, with practical MATLAB guidance.

Contribution

It introduces and compares three specific uncertainty and sensitivity analysis methods tailored for agent-based models, providing implementation details.

Findings

Provides detailed explanation of Consistency, Robustness, and Latin Hypercube analyses.

Includes MATLAB implementation guidance for each method.

Clarifies how to interpret results of these analyses in agent-based models.

Abstract

Multiscale, agent-based mathematical models of biological systems are often associated with model uncertainty and sensitivity to parameter perturbations. Here, three uncertainty and sensitivity analyses methods, that are suitable to use when working with agent-based models, are discussed. These methods are namely Consistency Analysis, Robustness Analysis and Latin Hypercube Analysis. This introductory review discusses origins, conventions, implementation and result interpretation of the aforementioned methods. Information on how to implement the discussed methods in MATLAB is included.