Agent-based modeling and life cycle dynamics of COVID-19-related online collective actions

Abstract

AbstractThe outbreak of COVID-19 has greatly threatened global public health and produced social problems, which includes relative online collective actions. Based on the life cycle law, focusing on the life cycle process of COVID-19 online collective actions, we carried out both macro-level analysis (big data mining) and micro-level behaviors (Agent-Based Modeling) on pandemic-related online collective actions. We collected 138 related online events with macro-level big data characteristics, and used Agent-Based Modeling to capture micro-level individual behaviors of netizens. We set two kinds of movable agents, Hots (events) and Netizens (individuals), which behave smartly and autonomously. Based on multiple simulations and parametric traversal, we obtained the optimal parameter solution. Under the optimal solutions, we repeated simulations by ten times, and took the mean values as robust outcomes. Simulation outcomes well match the real big data of life cycle trends, and validity and robustness can be achieved. According to multiple criteria (spans, peaks, ratios, and distributions), the fitness between simulations and real big data has been substantially supported. Therefore, our Agent-Based Modeling well grasps the micro-level mechanisms of real-world individuals (netizens), based on which we can predict individual behaviors of netizens and big data trends of specific online events. Based on our model, it is feasible to model, calculate, and even predict evolutionary dynamics and life cycles trends of online collective actions. It facilitates public administrations and social governance.