Particle swarm for calibration of land-use and transport integrated models

Abstract

Over the past few years, urban modelers have focused their efforts into making more elaborate simulations, which yield more explanatory power by merging together an increasing amount of urban features. In particular, the integration of Land-Use and Transport Interactions (LUTI) was a key step for urban models. The calibration of these increasingly complex simulations has become a very challenging task, making pure mathematical models inaccurate because of the many assumptions they rely on. This paper aims to extend the tools available to modelers in calibrating LUTI models, in order to manage this problem in a more efficient manner. This paper proposes a metaheuristic approach which considers the calibration problem as a global optimization problem, without relying on assumptions on the model. The chosen optimization algorithm is the Particle Swarm Optimization (PSO). It gradually improves the parameters of the model by testing and comparing the results with survey-based observations.