RANDIGRAPH · Random Directed Graph Models with Applications to Epidemic Processes and Communication Networks
FP6 — Marie Curie Actions (Human Resources and Mobility)
- Duration
- 2004-07-26 → 2005-07-25
- EU contribution
- €112,112
- Participants
- 1
- Scheme
- EIF
Lines connect the coordinator with its partners.
Results in brief
Final Activity Report Summary - RANDIGRAPH (Random directed graph models with applications to epidemic processes and communication networks)
Many complex biological and physical systems such as epidemic processes and communication networks (e.g. the internet) can be studied by developing a random directed graph model of the system. To capture the essential features of such systems it may be necessary to construct models with certain additional properties or constraints. The main result of the research carried out under this fellowship has been to show how to construct a new, flexible random mapping model which can be used to model inhomogeneous systems where some vertices (i.e. nodes) are more 'popular' than others. We have characterised the graphical structure of this new model under arbitrary assumptions about the underlying vertex 'popularity distribution' (i.e. vertex in-degree distribution), and applications of this model to the statistical modelling of epidemic processes have been developed. The component structure of random directed graphs, and graphs which are generated by them, under some structural restrictions or degree assumptions, which follow from applications in cluster analysis, cryptology and computer science have also been investigated.
Data: CORDIS, © European Union
Project objective
/ propose to investigate random directed graphs which incorporate a 'popularity measure' into their construction. We assign each vertex a random 'weight' which represents its relative 'attractiveness'. The simplest case leads to compound random mappings. It is expected that results for compound random mappings will contribute to the development of a theoretical framework for Bayesian inference for epidemic processes on contact networks. I also propose to investigate random models that incorporate arbitrary degree distributions. This provides another method of incorporating a 'popularity measure' into the model. By considering arbitrary degree distributions it is possible to mimic the highly skewed degree distributions seen in many networks such as the World Wide Web. Such models can be used for simulating and testing network algorithms. There are also connections between the analysis of epidemic processes and the analysis of algorithms in cryptology which may be elucidated by this investigation. To investigate the proposed non-uniform random digraph models I need to gain expertise in non-combinatorial techniques. I expect to acquire it through collaboration and interaction with the probability, applied statistics, and statistical physics groups at Heriot-Watt University. In addition, I would like to explore with the applied scientists at Heriot-Watt the potential applications of these random digraph models to a broad range of problems in both epidemic processes and communication networks. This training will enable me to return to my home institution to lead and direct a new multidisciplinary research group focusing on these applications of random digraphs. This project enhances European excellence in science by addressing the problem of fragmentation within the European Research Area and by fostering links between centres of excellence in discrete mathematics in Poznan and in probability, statistics and mathematical physics at Heriot-Watt.
Original text from CORDIS.
Participants
- HERIOT-WATT UNIVERSITY · EDINBURGHCoordinatorUnited Kingdom
Links
Data: CORDIS, © European Union