A stochastic model of randomly accelerated walkers for human mobility

Area of Science:

• Complex systems
• Statistical physics
• Transportation science

Background:

• Human mobility studies often use Lévy flight models for displacement patterns.
• Power law fits are commonly associated with scale-free superdiffusive random walks.
• Interpreting complex systems solely from data fits can lead to errors.

Purpose of the Study:

• To evaluate the applicability of the Lévy flight model to vehicle travel times and speeds.
• To develop a more accurate mechanistic model for human mobility.
• To investigate the underlying dynamics of vehicle movement patterns.

Main Methods:

• Analysis of a large dataset (780,000 private vehicles) of vehicle trajectories in Italy.
• Testing the Lévy flight model against empirical data for travel times and speeds.
• Introduction and validation of an accelerated random walk model with random acceleration kicks and exponentially decaying travel times.

Main Results:

• The Lévy flight model fails to explain the observed travel times and speeds of vehicles.
• The proposed accelerated random walk model accurately captures empirical observations.
• The model generates short-tailed distance distributions that can resemble truncated power laws.

Conclusions:

• Purely descriptive models like Lévy flights have limitations in explaining complex mobility dynamics.
• Accelerated random walks offer a mechanistic explanation for vehicle movement patterns.
• Understanding the underlying dynamics is crucial for accurate mobility modeling.