ABSTRACT
In this article, a spatiotemporal schedule model of agents’ behavior is established. With the main concern for the behavior of agent at the individual level, a hierarchical model is designed, in which activities are classified as mandatory activity, traveling activity, and flexible activity. Naturally, activities linked by time order constitute the one-day behavior pattern, which may differ on weekdays and at weekends. The one-week behavior pattern is composed of weekday behavior pattern and weekend behavior pattern. It is significant to make the model data driven when more data sources are available in the era of big data, so a hierarchical model configuration mechanism is proposed in this article. Then, an algorithm of activity chain generation is put forward. As a case study, taking student agent as an example, the activities chain generating results proved to be temporally heterogeneous at the individual level. We conclude that the proposed model may be effective in modeling agents’ behavior in an artificial society.
