CHAPTER 68

European Air- and Rail-Transport

Dominik Grether

This chapter discusses simulation of air- and rail-transport technology and passengers using

MATSim. There is no great difference in overall travel times between middle-range rail and air

transportation. Airports and railway stations are affected by capacity and opening time constraints.

For passengers and goods, geospatial location is an important property. Both modes, but especially

air transport, are faced with difficult capacity restrictions at certain departure times.

This chapter discusses how MATSim can be applied to capture these constraints and how in-

transport could also be applied to rail transport (Quick, 2012).

aspects requiring further study. The choice set generation and plans removal algorithm of MATSim

is discussed in detail; that is also the subject of Section 97.4. Modeling, results, and studies of this

chapter present the highlights of Grether (2014, Chapter 6, pp. 119), in more detail.

How to cite this book chapter:

Grether, D. 2016. European Air- and Rail-Transport. In: Horni, A, Nagel, K and Axhausen, K W.

(eds.) The Multi-Agent Transport Simulation MATSim, Pp. 419–428. London: Ubiquity Press.

Figure 68.1: Layout of airports in the air transport network: In- and outbound runways are mod-

eled by separate links connected by taxiways and a link representing the apron. There the transit

stop facility is attached.

Source: Grether et al. (2013)

destinations are only served on specific days.

used either for inbound or outbound flights, with taxiways connecting the runways to the apron.

terminate (Figure 68.1). Each airport pair is directly connected by airway links, one for each flight

and direction of travel (Figure 68.2). Maximum speed on any of these links is calculated based on

distance and flight duration provided by OAG. Times for taxi, take-off, and landing are also taken

into account, i.e., flight duration is reduced by the time needed from push-back to airborne before

air traffic network, is created for each flight offered by this airline. Mutual interferences of aircrafts

en-route are not included in the studies presented in this chapter.

Figure 68.3: European air network with country borders in the background (country borders

Source: Grether et al. (2013)

To represent individual aircraft in the simulation, transit vehicles are created on the basis of OAG

data. IATA aircraft codes, operating airlines, and seating capacities are reflected in the respective

aircraft representation for every flight. Information about boarding times, i.e., passenger flow per

evaluated by the utility function according to Section 3.2. A typical duration, ttyp,q, of 21 hours is

set for this activity type. The time virtual persons arrive at the origin airport and start waiting for

a connection is drawn randomly from a uniform distribution in 4 am to 6 pm, UTC. This reflects

estimated typical opening hours of European airports. No other time constraints are set, thus the

only incentive for virtual persons is to reduce overall travel time and maximize time spent at the

As a scenario for air transport technology, a coverage model from Europe to world wide destina-

tions is used; with the synthetic population, it serves as input for the simulation. The assignment of

flights to the desired O-Dconnection, i.e., the passenger routing, is calculated by MATSim’s default

public transit routing module.

shift their departure time randomly within a 2 hour interval. Another 10 % may seek a new route,

are switched off. For another 100 iterations only the MNL is used by passengers to select a plan.

iteration 600.