Information Technology Reference
In-Depth Information
is considered in our experiments but we plan to include the bottlenecks in the
future and the model can handle this by including some time offsets to allow for
the time reaching the bottlenecks.
The model of the gate allocation problem (GAP) which is presented here is
meant to be an early stage planning tool which attempts to resolve conflicts at
gates. Gates are divided into groups and the number of conflicting allocations is
limited within each of the groups. The groups which are visible in Figure 1 have
been introduced based upon the layout of the taxiways and our understanding
of the problem.
Gate constraints, which result from gate attributes, are also modelled. A gate
is an area adjacent to a terminal building where an aircraft can park so that
passengers can easily be loaded and unloaded. While an aircraft is on a gate all
necessary work such as fuelling, cleaning, baggage loading and ooading are also
performed. The work and time that an aircraft needs on a gate depends primarily
upon the airline that is operating it and the flight destination. There are also
areas which are not attached to terminals where aircraft can park, called remote
stands. When an aircraft is on a remote stand passengers are usually transported
to and from a terminal building by bus. The remote stands are usually used only
when all gates are occupied or if specifically requested by an airline. A number
of constraints result from the gate attributes, such as size restrictions and/or
shadowing restrictions (see Section 4.2). Some aircraft will need to be allocated
to different stands for their arrival and departure processes, for instance due to
different security requirements for the arriving and departing flights. These may
require a tug operation in between them to move the aircraft between stands.
Such aircraft can be modelled as two flights within this model, although the
experimental data used for these results does not detail such aircraft and these
are not the main focus of this work.
The GAP for one day (whole airport) can be very hard to solve using exact
methods. However, the whole airport problem can usually be decomposed by
terminals, since it is rare for an aircraft to be moved between terminals (all
passengers would have to change which terminal they checked in at) and airlines
usually have agreements with airports specifying which terminal (and sets of
gates) they will use. Allocating flights to a different terminal would be very
inconvenient for the airline because all of the necessary ground services would
have to be moved to a different part of the airport. However, synchronisation of
terminals could be needed later (when bottlenecks are included) since some of
taxiways are used to reach multiple terminals.
The proposed terminal based decomposition allowed the solution of instances
used in the experiments. But the solution process is slow due to the number
of possibilities that have to be checked. Similar problems have been reported
by other researchers working on the GAP [2,10]. It would be very hard to solve
larger instances for busy hub terminals to optimality, hence future work will con-
sider potential decompositions of the problem and/or heuristic approximations
to generate additional constraints and prune the search space.
