Facilitating Parameter Estimation and Sensitivity Analysis of Agent-Based Models: A Cookbook Using NetLogo and 'R'

AbstractAgent-based models (ABMs) are increasingly used in the management sciences. Though useful, ABMs are often critiqued: it is hard to discern why they produce the results they do and whether other assumptions would yield similar results. To help researchers address such critiques, we propose a systematic approach to conducting sensitivity analyses of ABMs. Our approach deals with a feature that can complicate sensitivity analyses: most ABMs include important non-parametric elements, while most sensitivity analysis methods are designed for parametric elements only. The approach moves from charting out the elements of an ABM through identifying the goal of the sensitivity analysis to specifying a method for the analysis. We focus on four common goals of sensitivity analysis: determining whether results are robust, which elements have the greatest impact on outcomes, how elements interact to shape outcomes, and which direction outcomes move when elements change. For the first three goals, we suggest a combination of randomized finite change indices calculation through a factorial design. For direction of change, we propose a modification of individual conditional expectation (ICE) plots to account for the stochastic nature of the ABM response. We illustrate our approach using the Garbage Can Model, a classic ABM that examines how organizations make decisions.

Download Full-text

Book Review: "Uncertainty and sensitivity analysis in archaeological computational modeling" edited by Marieka Briuwer Burg, Hans Peeters, William A. Lovis

10.31235/osf.io/v5du3 ◽

2017 ◽

Author(s):

Ben Marwick

Keyword(s):

Land Use ◽

Sensitivity Analysis ◽

Computational Modeling ◽

Cultural Transmission ◽

Agent Based Models ◽

Agent Based ◽

Sources Of Uncertainty ◽

Uncertainty And Sensitivity Analysis ◽

Archaeological Models

This volume is collection of papers emerging from a forum at the 2014 SAA meetings. The papers are motivated by the question of how we can measure and interpret uncertainty in quantitative archaeological models, specifically by using sensitivity analysis. The types of models discussed in this volume include geo-referenced models of past environments to infer hunter-gather land use, and agent-based models of cultural transmission processes. They explore various sources of uncertainty, and implement sensitivity analysis by assessing how the output of the models varies according to changes in the inputs. The motivation for this collection is the editors' observations that archaeologists lack a discipline-based protocol for testing models.

Download Full-text

Towards a Physiological Scale of Vocal Fold Agent-Based Models of Surgical Injury and Repair: Sensitivity Analysis, Calibration and Verification

Applied Sciences ◽

10.3390/app9152974 ◽

2019 ◽

Vol 9 (15) ◽

pp. 2974 ◽

Cited By ~ 3

Author(s):

Aman Garg ◽

Samson Yuen ◽

Nuttiiya Seekhao ◽

Grace Yu ◽

Jeannie Karwowski ◽

...

Keyword(s):

Sensitivity Analysis ◽

Confidence Intervals ◽

Empirical Data ◽

Vocal Fold ◽

Cell Populations ◽

Model Parameters ◽

Agent Based Models ◽

Agent Based ◽

Injury And Repair ◽

Surgical Injury

Agent based models (ABM) were developed to numerically simulate the biological response to surgical vocal fold injury and repair at the physiological level. This study aimed to improve the representation of existing ABM through a combination of empirical and computational experiments. Empirical data of vocal fold cell populations including neutrophils, macrophages and fibroblasts were obtained using flow cytometry up to four weeks following surgical injury. Random Forests were used as a sensitivity analysis method to identify model parameters that were most influential to ABM outputs. Statistical Parameter Optimization Tool for Python was used to calibrate those parameter values to match the ABM-simulation data with the corresponding empirical data from Day 1 to Day 5 following surgery. Model performance was evaluated by verifying if the empirical data fell within the 95% confidence intervals of ABM outputs of cell quantities at Day 7, Week 2 and Week 4. For Day 7, all empirical data were within the ABM output ranges. The trends of ABM-simulated cell populations were also qualitatively comparable to those of the empirical data beyond Day 7. Exact values, however, fell outside of the 95% statistical confidence intervals. Parameters related to fibroblast proliferation were indicative to the ABM-simulation of fibroblast dynamics in final stages of wound healing.

Download Full-text

Sensitivity analysis methodologies for analysing emergence using agent-based models

10.18174/423172 ◽

2017 ◽

Author(s):

Guus ten Broeke

Keyword(s):

Sensitivity Analysis ◽

Agent Based Models ◽

Agent Based

Download Full-text

Using multiple scale space-time patterns in variance-based global sensitivity analysis for spatially explicit agent-based models

Computers Environment and Urban Systems ◽

10.1016/j.compenvurbsys.2019.02.006 ◽

2019 ◽

Vol 75 ◽

pp. 170-183 ◽

Cited By ~ 5

Author(s):

Jeon-Young Kang ◽

Jared Aldstadt

Keyword(s):

Sensitivity Analysis ◽

Global Sensitivity Analysis ◽

Multiple Scale ◽

Scale Space ◽

Space Time ◽

Spatially Explicit ◽

Agent Based Models ◽

Agent Based ◽

Global Sensitivity ◽

Time Patterns

Download Full-text

Hybrid approach for parameter estimation in agent-based models

10.1101/175661 ◽

2017 ◽

Cited By ~ 3

Author(s):

Jill A. Gallaher ◽

Andrea Hawkins-Daarud ◽

Susan C. Massey ◽

Kristin R. Swanson ◽

Alexander R.A. Anderson

Keyword(s):

Cancer Research ◽

Parameter Estimation ◽

Hybrid Approach ◽

New Method ◽

Agent Based Model ◽

Agent Based Models ◽

Agent Based

AbstractAgent-based models are valuable in cancer research to show how different behaviors emerge from individual interactions between cells and their environment. However, calibrating such models can be difficult, especially if the parameters that govern the underlying interactions are hard to measure experimentally. Herein, we detail a new method to converge on parameter sets that fit an agent-based model to multiscale data using a model of glioblastoma as an example.

Download Full-text