Agent-based modeling of spreading infectious diseases: state-of-the-art

Agent-based simulation modeling provides additional opportunities to study the patterns of pathogen spread among populations, taking into account the complexity and stochasticity of the epidemic process. Agent-based modeling is considered as a computational approach in which agents with predefined characteristics can interact with each other and with the environment according to pre-specified rules. Here I consider the historical background of agent-based modeling in the field of infectious diseases, describe the basic definitions and classifications, and discuss strengths and weaknesses of agent-based modeling. The article details four interconnected main components that are subject to modeling: disease features (transmission routes, features of the infectious process), the population, movement patterns, and the environment. The article also addresses the need for validation of agent-based models. The reader's attention is drawn to the following important features of agent-based simulation models: the ability to model various scenarios on different scales (global, national, regional); the ability to use them in epidemiological studies when controlled experiments are impossible (e.g., consequences of non-compliance with preventive measures, spread of «cultural pathogens»); agents can make different decisions depending on their characteristics; consideration of behavioral aspects at the individual level; the ability to account for individual mobility and social contacts of agents. Agent-based simulation models are also well-suited for epidemiological modeling, particularly in the field of infectious disease surveillance, including emerging infections (e.g., COVID-19).