A Multi-model Algorithm for the Optimization of Congested Networks
C Osorio, M Bierlaire, Transport and Mobility Laboratory Ecole Polytechnique Fédérale de Lausanne, CH
The complexity of traffic flow dynamics in congested urban road networks has motivated the development of advanced traffic simulators that are able to capture in detail the interactions of individual drivers with both the network supply components and with adjacent drivers. These simulation tools can provide accurate network performance estimates in the context of scenario-based analysis or sensitivity analysis. Nevertheless, their integration within an optimization framework remains an intricate process. We believe that in order to perform both fast and reliable optimization for congested networks, information from the simulation tool should be combined with information from a surrogate network model that analytically captures the structure of the underlying problem.
We present a trust-region optimization framework (Carter (1986)) that combines the use of these two models, leading to a trade-off between numerical performance and accuracy. We illustrate the use of this framework by considering the problem of traffic signal optimization for a congested urban road network. The analytical model considered is a finite capacity queueing network model (Osorio and Bierlaire (2009)) that has been successfully used in past studies for fixed-time traffic signal control (Osorio and Bierlaire (2008)). The simulation tool used is a calibrated microscopic traffic simulation model of the Lausanne city center (Dumont and Bert (2006)). The most recent advances of this ongoing project will be presented and discussed.
- Carter, R. G. (1986). Multi-model algorithms for optimization, PhD thesis, Rice University.$
- Dumont, A. G. and Bert, E. (2006). Simulation de l?agglom´eration Lausannoise SIMLO, Technical report, Laboratoire des voies de circulation, ENAC, Ecole Polytechnique Fédérale de Lausanne.
- Osorio, C. and Bierlaire, M. (2008). Network performance optimization using a queueing model, Proceedings of the European Transport Conference (ETC), Noordwijkerhout, The Netherlands.
- Osorio, C. and Bierlaire, M. (2009). An analytic finite capacity queueing network model capturing the propagation of congestion and blocking, European Journal Of Operational Research 196(3): 996?1007.
Association for European Transport