Agent-Based Modeling in Economic Science: Theoretical and Methodological Foundations

This article provides a theoretical and methodological justification for agent-based modeling (ABM) as a tool for overcoming the crisis of orthodox economic theory. The authors conduct a systematic critical analysis of the limitations of econometrics and DSGE models as research tools in modern economic science under conditions of structural instability and agent heterogeneity. The paper presents the mathematical architecture of an agent-based computing system, including network structures of interactions, and conducts a comparative analysis of ABM, econometrics, and mathematical modeling. The authors substantiate the conclusion that ABM is not an alternative, but rather a necessary complement to traditional methods, allowing for the modeling of emergent macroeffects based on the microbehavior of agents. Particular attention is given to the prospects of methodological triangulation, combining econometric verification, analytical models, and simulation experiments for analyzing complex economic systems. In a comparative table and a separate section, the authors provide a detailed comparison of three research tools in modern economic science: econometrics, mathematical modeling, and agent-based modeling. Finally, the authors conclude that agent-based modeling, with its high flexibility and realism, is an effective method for scenario analysis and studying complex economic systems with multiple heterogeneous agents. However, it requires careful interpretation and triangulation of modeling results, taking into account the significant limitations and characteristics of agent-based modeling as a research tool.