Calibration of Traffic Simulations Using Simulated Annealing and GPS Navigation Records

A traffic simulation tool provides a virtual environment to efficiently analyze current traffic conditions and quickly measure the impact of changes to either transport infrastructure or driving rules. Realizing the full potential of traffic simulations depends on correct parameter setting. In this work, we propose a method to calibrate traffic simulations using available transportation data from Costa Rica. The data comes from Global Position System (GPS) navigation records that only show the traffic speed in different sectors. The calibration algorithm aims to solve the inverse problem of finding the actual traffic flows in all routes to accurately reproduce real traffic conditions. We managed to calibrate the simulations for four case studies and leveraged our program to design solutions that ease traffic conditions in those scenarios. The impact and applications of this work are plenty. First, additional calibration techniques can be explored. Second, available data for more general settings can be exploited. Third, our tool can be integrated as a useful resource for analysis and decision making in urban mobility studies.