Modelos Basados en Agentes (MBA): definición, alcances y limitaciones

This guide provides an introductory overview of Agent-Based Modeling (ABM), a 'bottom-up' simulation technique used to study complex systems. It defines the components of ABMs, explains their utility in capturing emergent behavior and heterogeneity, and offers guidance on when to apply the method, how to verify and validate models, and how to document them using the ODD protocol.

• Agent-Based Modeling (ABM) is a 'bottom-up' technique where a system is represented as a collection of autonomous decision-making entities called agents. Unlike traditional analytical methods that focus on system equilibrium (top-down), ABMs model the behavior and interactions of individual agents and local objects to generate system-level patterns.
• An ABM consists of three primary components: agents, an environment, and rules. Agents are physical or virtual entities (such as people, animals, or organizations) with their own resources, goals, and sensory capabilities. The environment is the virtual space where they interact, and the rules define the relationships and sequences of actions.
• ABMs are specifically suited for complex systems characterized by interdependencies, heterogeneity, and nested hierarchies. They are used to study 'emergence,' where global system properties are different from local properties and cannot be predicted simply by extrapolating the behavior of individual components.
• The guide identifies five specific scenarios where ABM is an appropriate approach: when there are complex interactions (non-linear or discrete), heterogeneous populations, complex topological interactions, complex behaviors (stochastic or non-definable by aggregate transition rates), and when spatial components are relevant to the system's functioning.
• Verification and validation are critical but challenging in ABMs. Verification ensures the model is built correctly through 'code walks,' testing sub-models, and running 'small world' simulations. Validation ensures the model represents the real system through actor involvement, comparison with historical data or theories, and model-to-model (M2M) comparisons.
• To ensure transparency and repeatability, the document strongly recommends using the ODD (Overview, Design concepts, and Details) protocol for documenting ABMs, as the method initially lacked a standard communication language.

APA

Cardoso, C., Bert, F., & Podestá, G. (n.d.). Modelos Basados en Agentes (MBA): definición, alcances y limitaciones. Inter-American Institute for Global Change Research. https://iai.int/admin/site/sites/default/files/uploads/2014/03/Cardoso_et_al_Manual_ABM.pdf

Chicago

Cardoso, Carolina, Federico Bert, and Guillermo Podestá. Modelos Basados en Agentes (MBA): definición, alcances y limitaciones. Inter-American Institute for Global Change Research, n.d. https://iai.int/admin/site/sites/default/files/uploads/2014/03/Cardoso_et_al_Manual_ABM.pdf.

Wikipedia

{{cite report |last1=Cardoso |first1=Carolina |last2=Bert |first2=Federico |last3=Podestá |first3=Guillermo |title=Modelos Basados en Agentes (MBA): definición, alcances y limitaciones |publisher=Inter-American Institute for Global Change Research |url=https://iai.int/admin/site/sites/default/files/uploads/2014/03/Cardoso_et_al_Manual_ABM.pdf |access-date=17 August 2026 |via=Climate Insights Directory}}

BibTeX

@techreport{cardosondmodelos, author = {Cardoso, Carolina and Bert, Federico and Podestá, Guillermo}, title = {{Modelos Basados en Agentes (MBA): definición, alcances y limitaciones}}, institution = {Inter-American Institute for Global Change Research}, url = {https://iai.int/admin/site/sites/default/files/uploads/2014/03/Cardoso_et_al_Manual_ABM.pdf}, urldate = {2026-08-17}, note = {Indexed by Climate Insights Directory} }