Agent-Based Modeling (ABM)
Parent article: ABM: The Method That Starts with the Individual
Agent
Computational entity with a set of behavioral rules, capable of perceiving its local environment and acting on it. An agent does not require perfect rationality or global information. May be reactive (responds to immediate stimuli), deliberative (plans over representations of the world), or hybrid.
Related: actor, node, individual, cellular automaton Cross-references: Robotics / Swarm Systems, Artificial Intelligence / Machine Learning Key work: Epstein, J. M., & Axtell, R. (1996). Growing Artificial Societies. Brookings Institution Press.
Methodological individualism
Principle that social phenomena must be explained from individual action, and are not reducible to properties of the collective independent of the individuals composing it. Origin: Carl Menger (1871). Implicit basis of ABM and explicit foundation of praxeology. ABM is the computational formalization of this principle.
Related: reductionism, praxeology, human action, aggregation Cross-references: Praxeology / Austrian Political Economy, Social Sciences / Behavioral Economics Key work: Menger, C. (1871/1950). Principles of Economics. Free Press.
Emergence
The aggregate result of interactions between agents that no individual rule specifies or can directly predict. In Schelling: segregation emerges from soft neighborhood preferences. In Axelrod: cooperation emerges from iterated interaction between self-interested agents. In Kiyotaki-Wright: money emerges from individual exchange decisions.
Related: emergent properties, complexity, self-organization Cross-references: Complex Systems / Network Science, Robotics / Swarm Systems Key work: Schelling, T. C. (1971). Dynamic models of segregation. Journal of Mathematical Sociology, 1(2), 143–186.
Validation
Process by which it is assessed whether an ABM reproduces observable patterns of the real system it represents. Methodologically distinct from statistical hypothesis testing: the model does not test a hypothesis about parameters; it generates patterns compared with empirical patterns. A model can be validated without being calibrated, and vice versa.
Related: calibration, verification, pattern-oriented modeling, empirical data Cross-references: Artificial Intelligence / Machine Learning Key work: Grimm, V., et al. (2005). Pattern-oriented modeling of agent-based complex systems. Science, 310(5750), 987–991.
Path dependence
Property of systems where final outcomes depend on the historical sequence of events, not only on current conditions or model parameters. In ABM: small differences in initial conditions or in the order of interactions can produce radically different final outcomes. Monetary emergence exhibits path dependence: which good becomes money depends on the order of early exchanges.
Related: non-linearity, bifurcation, lock-in, history Cross-references: Complex Systems / Network Science, Praxeology / Austrian Political Economy Key work: Arthur, W. B. (1994). Increasing Returns and Path Dependence in the Economy. University of Michigan Press.
Misesian apriorism
Methodological position holding that the principles of human action are known a priori, not through empirical observation, because action is both the knowing subject and cannot become its own object of study without changing its nature. In tension with ABM, which is experimental and falsifiable. Mises would accept ABM as illustration of theoretical propositions, not as a source of them.
Related: praxeology, empiricism, rationalism, deductive method Cross-references: Praxeology / Austrian Political Economy Key work: Mises, L. von (1949). Human Action: A Treatise on Economics. Yale University Press.
References
Axelrod, R. (1984). The Evolution of Cooperation. Basic Books.
Epstein, J. M., & Axtell, R. (1996). Growing Artificial Societies: Social Science from the Bottom Up. Brookings Institution Press.
Menger, C. (1871/1950). Principles of Economics (J. Dingwall & B. F. Hoselitz, Trans.). Free Press.
Mises, L. von (1949). Human Action: A Treatise on Economics. Yale University Press.
Schelling, T. C. (1971). Dynamic models of segregation. Journal of Mathematical Sociology, 1(2), 143–186.