Information Technology Reference
In-Depth Information
9.5
Conclusion
The organization of the air transport network represented here reveals an interesting
process of spatial integration. The networks are articulated around a central core,
consisting of principal hubs where the arrangement of “reticular territories” distin-
guishes the three organizing zonal sets of the world economy: the United States,
Europe and Asia. In the United States, structures are hierarchical, and the various
territories are perfectly integrated with each other. Europe is an example of an
airspace with very little organization and low levels of hierarchy: there is a multitude
of airports and relations at the same level without a single organizing influence, as
in the United States. One may assume that this phenomenon will attenuate over the
next few years with the new European transportation policy (in particular, through
the implementation of the White Paper on the “European transport policy by 2010:
the hour of choices”).
Attention should be given to the influence of Asian airspace. Indeed, Asia is the
area currently showing the greatest growth in air transport. It will be necessary to
closely follow the evolution of the Asian sector because one may assume that the
principal worldwide hubs group will soon integrate with airports such as Tokyo,
Hong Kong and Singapore.
Economies, markets and the internal reorganizations of airline companies are
factors supplementing these tendencies. Power transfer from states to private
companies is leaving airspace less divided and consequently more inclined toward
competitiveness logics. Thus, private strategies have become the dominant orga-
nizing influence for the “reticular territories” of air transport. The only regulatory
authorities in this system are the international air organizations (IATA, ICAO). With
air transport being one of the most competitive sectors in the world, one must expect
constant evolution of its networks and an increasingly complex organization of the
reticular territories.
From a methodological point of view, the comparison of clustering methods can
be classified into two types of final results:
results that include a large number of geographically proximal areas but do not
highlight the hubs in the network;
results that highlight the denser parts of the networks and the hubs. These
methods are interesting but represent the star shapes of the networks very poorly.
Repeating the clustering methods with different datasets shows that neither of
these types of methods is satisfying in terms of the processes that need to be
delineated: the methods do not operate on the same order or with similar weights.
Thus, these types of exploratory methods are limited and could be supplemented
in the future by more data-driven methods and the selection of different steps
depending on the initial assumptions. These types of data-driven methods were first
tested on commuter networks (see Chap.
11
) and could be automated to test these
methods on different datasets. Thus, the field of clustering is in progress and is open
to many future research directions.
Search WWH ::
Custom Search