An Extended Exploration and Preferential Return Model for Human Mobility Simulation at Individual and Collective Levels
Abstract
Human mobility models have prevalent applications in traffic management, urban planning, and disease prevention. Modeling human mobility combining both the statistical physical mechanics and geographical constrains is not fully investigated. In this research, we extend the exploration and preferential return (EPR) model by considering both spatial heterogeneity and distance decay. The extended model takes rank distance in the distance decay function, and is validated in four cities based on the population distribution and the trajectory data. Our model not only reproduces the statistics of human mobility at both the individual and collective levels with high accuracy, but also has a robust prediction at both high and low resolutions. The study demonstrates the potential of applications of the human mobility models aggregating the collective and individual levels factors, and sheds light on the trade-off between simple' mobility rules and complex' geographical environments.
Ximeng Cheng
Postdoctoral Researcher
My main research interests are spatial-temporal big data mining, explainable artificial intelligence (XAI), GeoAI, time series analysis, and social sensing.
