Home Others Agent-based modeling

Agent-based modeling




Web pages




Agent-based models are computer simulations with which the interactions between people, things, places and time are examined. These are stochastic models created from the bottom up, which means that individual agents (often people in epidemiology) are assigned certain attributes. The agents are programmed to behave and interact with other agents and the environment in certain ways. These interactions produce effects that may differ from the effects of individual active substances. Agent-based modeling differs from traditional, regression-based methods in that, like system dynamics modeling, it enables the investigation of complex systems that exhibit independence from individuals and feedback loops in causal mechanisms. It is not limited to observed data and can be used to model counterfactual or experiments that may be impossible or unethical in the real world. However, agent-based modeling is not without its limitations. The data parameters (such as the reproductive rate in infectious diseases) are often difficult to find in the literature. In addition, the validity of the model can be difficult to assess, especially when modeling unobserved associations. Overall, agent-based models provide an additional tool for assessing the effects of exposure on outcomes. It is particularly useful when relationships, reciprocity, and feedback loops are known or suspected, or when real-world experiments are not possible.


Textbooks & Chapters

Think about complexity , by Allen Downy. Not written for epidemiologists or healthcare professionals, but this excellent, readable book by Allen Downey explains and provides examples of many of the original theories and principles of complex adaptive systems and agent-based modeling, such as Thomas Schelling's Dynamic Models of Segregation, Stephen Wolfram's work on cellular Automata as well as fractals and game theory. Chapters 1, 2, 6, 7 and 10 are particularly relevant and interesting. Book is online for free.

Where medicine went wrong: rediscovering the path to complexity , by Bruce J West. Weltwissenschaftlicher Verlag; 1st edition (October 9, 2006).
Although this text does not deal specifically with agent-based models, this interesting book deals with the potential errors in our traditional modeling of many physiological processes and disease states; i.e., in an attempt to impose Gaussian normality on inherently more complex systems. Written for clinicians, but a background in physics is very helpful.

Methodical articles

Systems Science Methods in Healthcare: Dynamics, Networks and Agents
Journal: Annual Public Health Review
Year of publication: 2012

Social network analysis and agent-based modeling in social epidemiology
Journal: Epidemiological Perspectives & Innovations
Year of publication: 2012

Causal thinking and complex systems approaches in epidemiology

endemic vs epidemic vs pandemic

Author (s): S Galea, M Riddle and GA Kaplan.
Year of publication: 2010

Application item

Agent-Based Simulation Platforms: Review and Development Recommendations
Journal: Simulation
Year of publication: 2006

Modeling and analysis of the global epidemiology of avian influenza

Author: DM Rao, A Chernyakhovsky und V Rao
Release year: 2009

Modeling of targeted multilayered flu pandemic containment in the USA

Author (s): ME Halloran, N. Ferguson, S. Eubank, et al.
Release year: 2008

Conceptual approaches to investigate health disparities

Author (in): AV Diez Roux
Year of publication: 2012

Understanding Long-Term Diffusion Dynamics in the Prevalence of Sexual Initiation in Adolescents: An Initial Investigation Using Agent-Based Modeling

Author (s): MG Orr and CR Evans
Year of publication: 2011

The role of the subway in a flu epidemic: a simulation in New York City New

Autor (en): P. Cooley, S. Brown, J. Cajka et al.
Year of publication: 2011

A spatial agent-based model for simulating the daily walking of adults in a city

Author: Y Yang, AV Diez Roux, AH Auchincloss, DA Rodriguez und DG Brown
Year of publication: 2011

Use of simple agent-based modeling to inform and improve the accessibility of the neighborhood

Author: H. Badland, M. White, G. Macaulay, et al.
Year of publication: 2013


Web pages

ABM blog written by Jeff Schank (UC Davis Professor of Psychology). It is a clearinghouse for information including lists of researchers using ABM, conferences, and additional websites.

Online guide for those new to agent-based modeling in the social sciences, a website created by Robert Axelrod and Leigh Tesfatsion (Iowa State University).

A presentation by Bruce West to a group of clinicians summarizing some of the concepts in his book Where Medicine Went Wrong: Rediscovering the Path to Complexity.


There's a course on Agent-based modeling, offered as part of the Epidemiology and Population Based Health Summer Institute at Columbia (EPIC)

Open ABM - is part of Computational Modeling for SocioEcological Science (CoMSES Net), a network dedicated to supporting and expanding the development and use of computational modeling in the social and life sciences. Not specifically health related.

Models of Infectious Disease Agent Study (MIDAS) - Network of NIH research groups working on the development of infectious disease models. The website contains some information about educational opportunities.

Agent-based models for population health
Next offer: June 1st - July 1st, 2016
Software used: NetLogo

Agent-based modeling boot camp and incubator for health researchers

Host / Program: Massachusetts Institute of Technology
Software used: AnyLogic

Model thinking

Host / Program: Coursera / University of Michigan
Course format: Online
Software used: NetLogo

Interesting Articles