Agent-based simulation from anonymized data: An application to Lille metropolis

Azise Oumar Diallo (1, 2) , Arnaud Doniec (1, 2) , Guillaume Lozenguez (1, 2) , René Mandiau (3)

1 IMT Nord Europe - Ecole nationale supérieure Mines-Télécom Lille Douai
2 CERI SN - IMT Nord Europe - Centre for Digital Systems
3 LAMIH - Laboratoire d'Automatique, de Mécanique et d'Informatique industrielles et Humaines - UMR 8201

The availability of transportation scenarios from open data introduces new research motivations, such as route pricing and pollution. Unfortunately, these data are often anonymized to protect the privacy of the survey participants. In this paper, we present a generation of a mobility simulation from anonymized data based on a modal choice model. We show how to reconstruct missing information such as the municipalities of residence and activities and the coordinates of the origins/destinations (O/D) of trips necessary for the estimation of the model. This model integrates an inter-modal alternative mixing car and public transport (pt) for commutes through park and ride (PR) facilities. A first application of the simulation model was made on the perimeter of the European Metropolis of Lille (MEL) where the activities zones were blurred and the O/D coordinates of trips were withdrawn from the household trip survey. The preliminary results obtained show that the simulation is capable of reproducing the behaviors of trips performed by combining car and pt. Therefore, this work can serve as a basis for testing inter-modality incentive policies (designated by “car + pt”) such as intra-urban tools or the creation of PR facilities.

Mots clés: Agent-based model; Open-data; Discrete mode choice; Inter-modality; MATSim; MEL

Domaines: Système multi-agents [cs.MA]

Azise Oumar Diallo, Arnaud Doniec, Guillaume Lozenguez, René Mandiau. Agent-based simulation from anonymized data: An application to Lille metropolis. The 12th International Conference on Ambient Systems, Networks and Technologies, Mar 2021, Varsovie, Poland. pp.164-171, ⟨10.1016/j.procs.2021.03.027⟩. ⟨hal-03380493⟩