Agent-based Modeling

by

Atsushi Nara

• LAST REVIEWED: 22 September 2021
• LAST MODIFIED: 22 September 2021
• DOI: 10.1093/obo/9780199874002-0236

Introduction

Agent-based modeling (ABM) has become widely accepted as a methodological tool to model and simulate dynamic processes of geographical phenomena. A growing number of ABM studies across a variety of domains and disciplines is partially explained by the development of agent-modeling tools and platforms, the availability of micro-data, and the advancement in computer technology and cyberinfrastructure. In addition to these technical reasons, another key motivation underlying ABM research is to address challenges embedded in conventional modeling approaches being relatively coarse, aggregate, static, normative, and inflexible across scales with a reductionist viewpoint (Batty 2005 cited under Application: Urban Systems.” With complexity science, including complex systems, complex adaptive systems, and artificial life, providing theoretical foundations and rationales, ABM is a computational methodology for simulating dynamic processes of nature and human systems driven by disaggregated, heterogeneous, and autonomous entities, i.e., agents, that interact among themselves and their environments. A key fundamental concept of the ABM framework is that a system emerges from the dynamic individual-level interactions from bottom-up, where the simulated outcome is more than the sum of its components. This bottom-up approach enables ABM to exhibit complex system dynamics, properties of which could include feedback effect, path-dependence, phase shift, non-linearity, adaptation, self-organization, tipping points, and emergence. Three key components of ABM are agents, their environment, and their decision rules. Agents are the crucial component in ABM where each individual agent has its own characteristics and agenda, assesses its surrounded situation, and makes decisions. Agents reside in an environment, which can represent a geographic space in case for spatially explicit agent-based models. Agents’ behavioral decisions and interactions within their environment are defined based on a set of rules, which can alter their status and location over time. The purpose of ABM research can be classified into theoretical exploration and empirical investigation as well as the combination of two. In the latter case, ABM can be used as an artificial laboratory experiment to explore what-if scenarios and to investigate how changes in agents, environments, and/or rules affect the macro-level outcomes. ABM has been applied to represent a wide variety of geographic processes and behaviors including but not limited to urban system, land-use/land-cover change, ecology, transportation, animal/human movement, behavioral geography, spatial cognition, transportation, and disease epidemiology. While the growing interest in ABM as a modeling methodology to simulate complex systems is remarkable, there exist various conceptual, methodological, and technical challenges.

General Overviews

Agent-based modeling (ABM) is a methodological tool to model and simulate complex adaptive systems with its emphasis on agents’ characteristics, decision-making behaviors, actions, reactions, and interactions at a local scale, which give rise to emergent system dynamics at a macro scale. Three essential components of ABM are agents, the environment, and decision rules. First, agents are the crucial component in ABM with their roots from artificial intelligence (AI). While there exist variations in its definition, researchers commonly list essential properties of agents as autonomous, heterogeneous, proactive, perceptive, communicative, and/or adaptive characteristics. Respectively these properties describe that each agent acts on its own without outside control, is individually distinct possessing heterogeneous characteristics, pursue its own behavioral goals, can sense other agents and its environment in its surroundings, interacts with other agents for its behavioral goals, and/or can change its behaviors by learning from its experience. Second, agents are situated within an environment, where they act and interact with other agents. An environment can be an aspatial model where the agents’ location is not a question of interest. For spatially explicit ABM, a geographic space can be represented in a form of vector objects (i.e., points, lines, polygons), raster grids, a hybrid of vector and raster objects, or graphical networks. Agent’s interactions are modeled with topological relationships describing agent’s connectedness and agent’s neighborhood where an agent can interact with others and its environment located nearby. Third, agents’ behavioral decisions and interactions within their environment are defined based on a set of rules, which drive evolving system dynamics over time. The decision-making rules, also known as behavioral or transition rules, are used to formulate the logic of agent’s behavioral processes that can alter its status over time relative to other agents and the environment. Behavioral rules can explicitly define the bounded rationality assumption that agents do not possess global information and act based on locally available information. Spatial ABM typically incorporates certain movement rules to change agent’s location in the environment. Researchers adopt ABM as a modeling approach for both theoretical exploration and empirical investigation of complex adaptive systems as well as the combination of two. Theoretical inquiry entails running controlled experiments to explore the implications of different theoretical frameworks or to develop new theories, whereas empirical models focus more on using actual data to simulate real-world phenomena (Manson 2007 and O’Sullivan, et al. 2016, cited under Key Challenges). Bonabeau 2002 reviewed and classified the benefits of ABM for simulating human systems. Epstein 2006 provided the rationale of ABM through the lens of generative social science and an epistemological perspective of ABM by discussing generative social science relative to deductive and inductive science. Heppenstall, et al. 2012 gave a comprehensive overview of ABM of geographical systems covering fundamental and conceptual works, techniques and methodologies, applications, and grand challenges. Macal and North 2010 presented an ABM tutorial in regards to key concepts, methodological foundations, software platforms and toolkits, and applications, while Macal 2016 discussed the background, definitions, challenges, and future directions of ABM. Manson, et al. 2012 discussed the profound relationship between ABM and complexity theory, which provided theoretical rationales for adopting ABM. A review article by O’Sullivan 2008 defined agents and ABM and provided implications of ABM for geographic information science. A review article by Torrens 2010 discussed the development and the use of ABM in physical and human geography and its research trends. Sengupta and Sieber 2007 overviewed roots of geospatial agents and categorized geospatial agents in ABM as artificial life geospatial agents (ALGA). From the point of view of AI, Wooldridge and Jennings 1995 discussed theoretical and practical issues associated with the design and construction of agents.

• Bonabeau, E. “Agent-Based Modeling: Methods and Techniques for Simulating Human Systems.” Proceedings of the National Academy of Sciences 99.Suppl. 3 (2002): 7280–7287.

  DOI: 10.1073/pnas.082080899

  Three ABM benefits discussed are its capability to capture emergent phenomena and to represent a human system composed of behavioral entities and its flexibility to incorporate multiple dimensions of agents’ characteristics, behaviors, and their interactions. Discussed four application areas, flows (e.g., evacuation, traffic), markets (e.g., stock, strategic simulation), organizational risk and design, and diffusion of innovation and adoptive dynamics.

• Epstein, J. M. Generative Social Science: Studies in Agent-Based Computational Modeling. Princeton, NJ: Princeton University Press, 2006.

  Discussed ABM with the perspective of generative social science with example models from archaeology to economics to epidemiology that illustrate foundational, theoretical, and empirical research aspects.

• Heppenstall, A., A. Crooks, L. M. See, and M. Batty, eds. Agent-Based Models of Geographical Systems. Berlin: Springer Netherlands, 2012.

  DOI: 10.1007/978-90-481-8927-4

  An essential source for researchers, students, and practitioners who apply ABM for simulating geographical processes.

• Macal, C. M. “Everything You Need to Know about Agent-Based Modelling and Simulation.” Journal of Simulation 10.2 (2016): 144–156.

  DOI: 10.1057/jos.2016.7

  Provided key ABM resources, publications, and communities. Discussed the distinct characteristics of ABM as compared to other standard simulation techniques such as discrete event simulation, system dynamics, Monte Carlo simulation, and continuous simulation.

• Macal, C. M., and M. J. North. “Tutorial on Agent-Based Modelling and Simulation.” Journal of Simulation 4.3 (2010): 151–162.

  DOI: 10.1057/jos.2010.3

  Introduced an overview ABM and discusses applications across a variety of disciplines and provided a series of questions that can help designing an agent-based model.

• Manson, S. M., S. Sun, and D. Bonsal. “Agent-Based Modeling and Complexity.” In Agent-Based Models of Geographical Systems. Edited by A. J. Heppenstall, A. T. Crooks, L. M. See, and M. Batty, 125–139. Berlin: Springer Netherlands, 2012.

  DOI: 10.1007/978-90-481-8927-4_7

  Provided theoretical rationales for adopting ABM and highlights five key issues underlying the intersection of ABM and complexity, which are (1) tensions between theoretical and empirical research; (2) calibration, verification, and validation; (3) scale; (4) equilibrium and change; and (5) decision making.

• O’Sullivan, D. “Geographical Information Science: Agent-Based Models.” Progress in Human Geography 32.4 (2008): 541–550.

  DOI: 10.1177/0309132507086879

  Covered example applications ranged from simple abstract models for thought experiments to spatially explicit models and detailed realistic simulations, and provided implications of ABM for geographic information science.

• Sengupta, R., and R. Sieber. “Geospatial Agents, Agents Everywhere . . . .” Transactions in GIS 11.4 (2007): 483–506.

  DOI: 10.1111/j.1467-9671.2007.01057.x

  Defined ALGA and provided distinctions between AI agents and geospatial agents. Discussed ALGA’s main themes and key characteristics at the operational level.

• Torrens, P. M. “Agent-Based Models and the Spatial Sciences.” Geography Compass 4.5 (2010): 428–448.

  DOI: 10.1111/j.1749-8198.2009.00311.x

  A review article of ABM in the spatial sciences providing a rich source of references, application studies in physical and human geography, and development environment for ABM.

• Wooldridge, M., and N. R. Jennings. “Intelligent Agents: Theory and Practice.” The Knowledge Engineering Review 10.2 (1995): 115–152.

  DOI: 10.1017/S0269888900008122

  Discussed the definition of agents, representations, and reasonings of agent’s properties, and designing and implementing agents based on agent theories from the point of view of AI.