The agent-based micro-simulation modelling technique for transportation planning is rapidly developing and is being applied to practice in recent years. In contrast to conventional four-step modelling with static assignment theory, this emerging technique employs a dynamic assignment principle. Based on summary of various types of traffic assignment models and algorithms, the thesis elucidates in detail the theories of two models, MATSim and EMME/2, which represent two genres of traffic assignment, i.e., dynamic stochastic stationary state assignment and static deterministic user equilibrium assignment. In the study, the two models are compared and validated to reflect both spatial and temporal variation of the traffic flow pattern. The comparison results indicate that numerical outputs produced by MATSim are not only compatible to those by EMME/2 but more realistic from a temporal point of view. Therefore, agent-based micro-simulation models reflect a promising direction of next generation of transportation planning models.
| Identifer | oai:union.ndltd.org:TORONTO/oai:tspace.library.utoronto.ca:1807/17506 |
| Date | 07 August 2009 |
| Creators | Gao, Wenli |
| Contributors | Miller, Eric |
| Source Sets | University of Toronto |
| Language | en_ca |
| Detected Language | English |
| Type | Thesis |
Page generated in 0.0016 seconds