Multiple Bayesian Models for the Sustainable City: The Case of Urban Sprawl

Giovanni Fusco (1) , Andrea G. B. Tettamanzi (2)

1 ESPACE - Études des Structures, des Processus d’Adaptation et des Changements de l’Espace
2 WIMMICS - Web-Instrumented Man-Machine Interactions, Communities and Semantics

Several possible models of urban sprawl are developed as Bayesian networks and evaluated in the light of available evidence, also considering the possibility that further, yet unknown models could offer better explanations. A simple heuristic is proposed in order to attribute a likelihood value for the unknown models. The case study of Grenoble (France) is then used to review beliefs in the different model options. The multiple models framework proves particularly interesting for geographers and planners having little available evidence and heavily relying on prior beliefs. This last condition is very frequent in research on sustainable cities. Further options of multiple models evaluations are finally proposed.

Mots clés: Grenoble; Bayesian Networks; Belief Revision; Model Selection; Uncertainty; Urban Sprawl

Domaines: Intelligence artificielle [cs.AI]; Géographie

Giovanni Fusco, Andrea G. B. Tettamanzi. Multiple Bayesian Models for the Sustainable City: The Case of Urban Sprawl. ICCSA 2017 - 17th International Conference on Computational Science and Its Applications, Jul 2017, Trieste, Italy. pp.392-407, ⟨10.1007/978-3-319-62401-3_29⟩. ⟨hal-01567726⟩