A design system for scalable agent-based models with multi-stage interactions of agents forming social connections (Q2206956)

This paper studies the ways to create an artificial society that serve as a testing ground for delivering socio-economic experiments and assessing the outcomes. The requirements of the model are determined as scalability, implementation of its parallelization, new units and additional processes flexibility, and adaptability for modeling other large socio-economic objects. Five methods are adopted here. Firstly, the system is organized as a base unit. Secondly, messaging between agents is separated from modules designed to simulate processes. Thirdly, the system supports the binding of agents to the place of residence in territory of the simulated country. Fourthly, the system supports the disappearance of agents and the emergence of new ones. Finally, the scalable model is developed as a separate executable module.