Analyzing time-varying trip distributions with a random-effect spatial OD dependence model

Area of Science:

• Transportation Science
• Spatial Analysis
• Econometrics

Background:

• Origin-destination (OD) trip distribution models are crucial for transportation planning.
• Existing spatial OD models often overlook unobservable zonal heterogeneity and random effects.
• Dynamic analysis of trip distributions requires advanced modeling techniques.

Purpose of the Study:

• To propose a novel random-effect spatial OD dependence model for time-varying trip distributions.
• To develop an advanced estimation method for simultaneously capturing spatial dependencies and zonal heterogeneity.
• To provide a robust tool for analyzing OD travel flow dynamics and informing transportation policy.

Main Methods:

• Development of a random-effect spatial OD dependence model.
• Application of maximum likelihood estimation with spectral decomposition.
• Validation through numerical experiments and analysis of real-world cellular signaling data.

Main Results:

• Successful simultaneous estimation of spatial dependences and unobservable zonal heterogeneity.
• Improved model fitting and parameter estimation accuracy.
• Empirical analysis reveals key insights into spatial effects and zonal variations in travel behavior.

Conclusions:

• The proposed model effectively captures dynamic spatial OD relationships and unobserved zonal factors.
• The estimation method enhances the precision of parameter identification in spatial models.
• This research offers a valuable framework for data-driven transportation policy formulation.