Summary
The goal of the PROXNET project, which will be realised during the mobility of Christophe Crespelle in the Algorithms group
of Bergen, is to open a new way for analysing, modelling and managing complex networks, through graph editing problems.
The reason why these networks are said to be complex is that they are loosely structured, due to the part of uncertainty and
randomness they contain. On the other hand, the real-world context where they come from strongly constrains their
organisation and gives them some specific structure. The difficulty in retrieving this structure is that it is altered by the noise
resulting from the uncertainty and randomness that these networks contain. In the PROXNET project, we will retrieve the
hidden structures of complex networks thanks to graph editing problems, which consist in changing some adjacencies of the
graph in order to obtain a desired property. We will develop the algorithms necessary to solve graph editing problems
relevant to our approach on huge instances of graphs, we will apply them to real-world datasets and use the results obtained
in order to design new models of complex networks.
One key theoretic challenge in this context is to optimally solve the editing problems considered, which are computationally
difficult (NP-hard). A very promising technique to do so is preprocessing. It consists in reducing the size of the instance by
applying some reduction rules, in such a way that the optimal solution on the reduced instance is the same as the one on the
original instance. The Algorithms group of the University of bergen is a world-wide leader on this topic which will constitute
the main training action of Christophe Crespelle during his mobility.
of Bergen, is to open a new way for analysing, modelling and managing complex networks, through graph editing problems.
The reason why these networks are said to be complex is that they are loosely structured, due to the part of uncertainty and
randomness they contain. On the other hand, the real-world context where they come from strongly constrains their
organisation and gives them some specific structure. The difficulty in retrieving this structure is that it is altered by the noise
resulting from the uncertainty and randomness that these networks contain. In the PROXNET project, we will retrieve the
hidden structures of complex networks thanks to graph editing problems, which consist in changing some adjacencies of the
graph in order to obtain a desired property. We will develop the algorithms necessary to solve graph editing problems
relevant to our approach on huge instances of graphs, we will apply them to real-world datasets and use the results obtained
in order to design new models of complex networks.
One key theoretic challenge in this context is to optimally solve the editing problems considered, which are computationally
difficult (NP-hard). A very promising technique to do so is preprocessing. It consists in reducing the size of the instance by
applying some reduction rules, in such a way that the optimal solution on the reduced instance is the same as the one on the
original instance. The Algorithms group of the University of bergen is a world-wide leader on this topic which will constitute
the main training action of Christophe Crespelle during his mobility.
Unfold all
/
Fold all
More information & hyperlinks
| Web resources: | https://cordis.europa.eu/project/id/749022 |
| Start date: | 01-02-2018 |
| End date: | 31-01-2020 |
| Total budget - Public funding: | 208 400,40 Euro - 208 400,00 Euro |
Cordis data
Original description
The goal of the PROXNET project, which will be realised during the mobility of Christophe Crespelle in the Algorithms groupof Bergen, is to open a new way for analysing, modelling and managing complex networks, through graph editing problems.
The reason why these networks are said to be complex is that they are loosely structured, due to the part of uncertainty and
randomness they contain. On the other hand, the real-world context where they come from strongly constrains their
organisation and gives them some specific structure. The difficulty in retrieving this structure is that it is altered by the noise
resulting from the uncertainty and randomness that these networks contain. In the PROXNET project, we will retrieve the
hidden structures of complex networks thanks to graph editing problems, which consist in changing some adjacencies of the
graph in order to obtain a desired property. We will develop the algorithms necessary to solve graph editing problems
relevant to our approach on huge instances of graphs, we will apply them to real-world datasets and use the results obtained
in order to design new models of complex networks.
One key theoretic challenge in this context is to optimally solve the editing problems considered, which are computationally
difficult (NP-hard). A very promising technique to do so is preprocessing. It consists in reducing the size of the instance by
applying some reduction rules, in such a way that the optimal solution on the reduced instance is the same as the one on the
original instance. The Algorithms group of the University of bergen is a world-wide leader on this topic which will constitute
the main training action of Christophe Crespelle during his mobility.
Status
CLOSEDCall topic
MSCA-IF-2016Update Date
28-04-2024
Images
No images available.
Geographical location(s)
