3.81 - Active Learning Metamodels for Scenario Discovery in Land-Use/ Transportation Simulators

Project Description

In this research, we are interested in combining active learning algorithms with simulation metamodeling approaches to study, understand and predict the behavior of simulation models, with a special focus on transport simulation models which can turn out to be overwhelming complex. Therefore, the exploration of their simulation input spaces can prove to be a tedious task. To this end, we aim to propose a novel active learning metamodeling methodology to address such drawbacks. This methodology should significantly reduce the computational workload often associated to such kind of simulators.

Research Team

CITTA

  • Francisco Antunes

University of Coimbra

  • Bernardete Ribeiro

Technical University of Denmark

  • Francisco Câmara Pereira

MIT

  • P. Christopher Zegras
Financial Support
  • FCT
Stage of Progress
  • Started in 2017