Post doctoral subject : Distributed Algorithms and Theorie of Belief Functions

Laboratory Heudiasyc UMR CNRS 7253
University of Technology of Compiègne
60200 Compiègne, France Heudiasyc


Postdoc Advisors
Prof. Bertrand Ducourthial (advisor)
+33 (0)3 44 23 46 46
Véronique Cherfaoui, (MC HDR, co advisor)
+33 (0)3 44 23 44 87


Context of the study
Distributed algorithms run on networks of computing and communicating units. Each unit participates to the global goal achievement, thanks to its local computations and to the communications to which it takes part with units of its neighborhood.

Such a distributed system can be disturbed by at least the following problems :

 units moves leading to unstable neighborhood and dynamic topologies (dynamic networks),
 failure of an entity that can break down (permanently or intermittently), or have an unpredictable behavior,
 failure of communication links in which messages can appear, disappear or be changed,
 deliberate attack from malicious agents trying to corrupt entities or messages to alter the system behavior in their neighborhood, and beyond.

Various algorithmic techniques have been proposed (or are under consideration) to counteract these disturbances, thereby ensuring that the distributed system meets its specifications : self-stabilization, failure detectors, redundancy, encryption, best-effort algorithms [1]...

However recent works focused on the use of the Theory of Belief Functions in networks [2] [3]. These preliminary works show a new approach to distributed algorithms because the tools of the Theory of Belief Functions can bring a lot in the diagnosis and control of the distributed system on the one hand and the design of algorithms resistant to disturbances on the other.

[1] Best-effort Group Service in Dynamic Networks B. Ducourthial, S. Khalfallah, F. Petit. 22nd ACM Symposium on Parallelism in Algorithms and Architectures, June 2010, Greece

[2] Distributed Data fusion for detecting Sybil attacks in VANETs. N. El Zoghby, V. Cherfaoui, B. Ducourthial, T. Denœux. Second International Conference on Belief Functions BELIEF 2012, Compiègne, France, 2012.

[3] Self-stabilizing distributed data fusion. B. Ducourthial, V. Cherfaoui, T. Denoeux. Soumis.


Postdoc description
This study is a continuation of previous work done in Heudiasyc laboratory regarding distributed data fusion. However, efforts will focus on the area of distributed systems as a subject of study and we will focus on the contribution of the theory of belief fonctions in this area, especially for distributed data management. The candidate will develop innovative distributed algorithms based on the Theory of Belief Function.


Possibly, depending on the candidate’s profile, fundamental applications would be considered for vehicular networks and fleets of UAVs, based on hardware and software platforms and the technical staff of the laboratory [4,5,6].

[4] http://www2.hds.utc.fr/pacpus

[5] https://www.hds.utc.fr/airplug

[6] http://www2.hds.utc.fr/uav-horus


Candidate’s profile
PhD in distributed systems, preferably.
Read, written, spoken English.
It is not necessary to speak French.


Documents required to apply
send to Bertrand.Ducourthial@utc.fr
• Curriculum vitae and motivation letter
• Last diploma
• Letter(s) of recommendation