Ab 2. November 2020 findet das Herbstsemester 2020 online statt. Ausnahmen: Veranstaltungen, die nur mit Präsenz vor Ort durchführbar sind.
Bitte beachten Sie die per E-Mail kommunizierten Informationen der Dozierenden.

101-0491-00L  Agent Based Modeling in Transportation

SemesterHerbstsemester 2018
DozierendeT. J. P. Dubernet, M. Balac
Periodizitätjährlich wiederkehrende Veranstaltung

KurzbeschreibungThis lectures provides a round tour of agent based models for transportation policy analysis. First, it introduces statistical methods to combine heterogeneous data sources in a usable representation of the population. Then, agent based models are described in details, and applied in a case study.
LernzielAt the end of the course, the students should:
- be aware of the various data sources available for mobility behavior analysis
- be able to combine those data sources in a coherent representation of the transportation demand
- understand what agent based models are, when they are useful, and when they are not
- have working knowledge of the MATSim software, and be able to independently evaluate a transportation problem using it
InhaltThis lecture provides a complete introduction to agent based models for transportation policy analysis. Two important topics are covered:

1) Combination of heterogeneous data sources to produce a representation of the transport system

At the center of agent based models and other transport analyses is the synthetic population, a statistically realistic representation of the population and their transport needs.
This part will present the most common types of data sources and statistical methods to generate such a population.

2) Use of Agent-Based methods to evaluate transport policies

The second part will introduce the agent based paradigm in details, including tradeoffs compared to state-of-practice methods.

An important part of the grade will come from a policy analysis to carry with the MATSim open-source software, which is developed at ETH Zurich and TU Berlin and gets used more and more by practitioners, notably the Swiss rail operator SBB.
LiteraturAgent-based modeling in general
Helbing, D (2012) Social Self-Organization, Understanding Complex Systems, Springer, Berlin.
Heppenstall, A., A. T. Crooks, L. M. See and M. Batty (2012) Agent-Based Models of Geographical Systems, Springer, Dordrecht.


Horni, A., K. Nagel and K.W. Axhausen (eds.) (2016) The Multi-Agent Transport Simulation MATSim, Ubiquity, London

Additional relevant readings, mostly scientific articles, will be recommended throughout the course.
Voraussetzungen / BesonderesThere are no strict preconditions in terms of which lectures the students should have previously attended. However, knowledge of basic statistical theory is expected, and experience with high-level programming languages (Java, R, Python...) is useful.