The Coq mailing list

[Fr�d�ric Gava <gava AT univ-paris12.fr>]

Dear Coq-club,

sorry for the noise, here the phd subject in text:

Doctoral research project:
Parallel verification of a high-level Petri net algebra
=======================================================

Keywords: verification, modularity, high-level parallelism, Petri nets

I. Host laboratory, funding and contacts
----------------------------------------

Funding: Paris regional doctoral scholarship, value of approximately
1570 Euros/month for a duration of 3 years.

Laboratory: Laboratory for Algorithmics, Complexity and Logic (LACL),
University Paris-12
 a. Director: Pr Gaëtan Hains
 b. Email: 
gaetan AT hains.org
 c. Team: ``communicating systems'' directed by Pr Elisabeth Pelz
 d. Postal address: LACL, Université Paris XII--Val de Marne,
                    UFR de Sciences et Technologie, Bâtiment P2, 2ème étage
                    61 avenue du Général de Gaulle, 94010 Créteil
 e. Phone: +33 (0)1 45 17 65 95
 f. Fax: +33 (0)1 45 17 66 01
 g. Web site: http://www.univ-paris12.fr/lacl/

Doctoral advisor: Pr Gaëtan Hains

Co-advisors:
 - Dr Frédéric Gava (coordinator of the VEHICULAR project)
    a. Position: Assistant Professor at LACL
    b. Phone : +33 (0)1 45 17 65 67
    c. Email: 
gava AT univ-paris12.fr
    d. Web page: http://www.univ-paris12.fr/lacl/gava/
 - Dr Franck Pommereau
    a. Position: Assistant Professor at LACL
    b. Phone: +33 (0)1 45 17 66 00
    c. Email: 
pommereau AT univ-paris12.fr
    d. Web page: http://www.univ-paris12.fr/lacl/pommereau/

Application: Send a CV and cover letter to the one of the
above-mentioned advisors (in PDF format preferably).

II. Doctoral research subject
-----------------------------

  The automatic verification of models (also called model-checking
[13] ) is a useful technique for automatic software verification, but
it is generally expensive in terms of memory capacity and computing
time (one speaks of its ``state-space explosion problem''). Many works
in this field have dealt with the acceleration of computation and
reduction of the space exploration. Parallelization is one of the
possible techniques to this end. However, it is in generally
implemented with low-level models (both for its specification as
model-checker and as parallel program [5] [8] [11] ) and with a more
or less naive data distribution (such as for example
http://quasar.cnam.fr/).

  The proposed work is first of all the design of a BSP [15] [16]
parallel algorithm for the construction of the exploration of the
state-space of a high-level coloured Petri net algebra, called M-Net
[7] (or of a clearly definite subset [3] [17] ) as well as a parallel
checking of logical properties on this graph (some logics like LTL
[13] or others are often very expressive but an expressive and
effectively verifiable subset will be necessary). The use of such a
strongly structured model will enable us to determine some structural
characteristics opening the way of a high-level parallelism (more
structured) and thus more efficient (and portable) which is the BSP
model.

  Then, a modular and polymorphic implementation (data type
independent, therefore ideal for the model-cheking of symbolic system
[2] and for a high-level algebra) will be carried out with a library
for high-level parallel programming, called BSMLlib [9] BSMLlib is
based on the OCaml (http://caml.inria.fr) language [14] and wad
developed jointly by LACL and LIFO (Computer Science laboratory of the
University of Orleans). Optimizations to the algorithm will be made in
particular with the extension of techniques of load-balancing based on
BSP costs [1] or on the structure of the logical formulae with respect
to the model [12] .

  Finally, tests of our software applied to computer security problems
(federating topic for LACL research) and to properties of high
performance programs (C, Ada or FORTRAN with a parallel computation
library such as MPI or with algorithmic skeletons) will be carried out
on the various PC's clusters at LACL and LIFO. (An algorithmic
skeleton is a function which can be implemented in parallel: each kind
of parallelism, divide-and-conquer, pipeline, task farming, etc., is
realised by a skeleton; this makes possible the easy and safe
parallelisation of programs, by using suitable skeletons that closely
reflect the parallelism intrinsic in the computational problem [4].)
Adaptations and additions will thereafter (or progressively) be added
to model and check logical properties of computer science's or
biological problems [18] which will be encountered in the literature.

  More details (in French) on the VEHICULAR project can be found at:
   http://www.univ-paris12.fr/lacl/gava/vehiculaire/vehiculaire.pdf

III. Bibliography
-----------------

[1] M. Bamha and G. Hains. An Efficient equi-semi-join Algorithm for
      Distributed Architectures. In V. S. Sunderam, G. Dick van
      Albada, P. M. A. Sloot, and J. Dongarra, editors, International
      Conference on Computational Science (ICCS 2005), Part II, number
      3515 in LNCS, Springer-Verlag, 2005.
[2] B. Boigelot. Symbolic Methods for Exploring Infinite State Spaces.
      PhD Thesis, volume 189, Collection des Publications de la
      Faculté des Sciences Appliquées de l'Université de
      Liège, 1999.
[3] R. Bouroulet, H. Klaudel, and E. Pelz. A semantics of security
      protocol language using a class of composable high-level petri
      nets. In Application of Concurrency to System Design, volume 4th
      ACSD, pages 99-108. IIIE, 2004.
[4] Jan Duennweber, Sergei Gorlatch, Anne Benoit and Murray Cole
      Integrating MPI-Skeletons with Web Service in Proceedings of
      ParCo 2005.
[5] H. Garavel, R. Mateescu, C. Joubert and al. DISTRIBUTOR and
      BCG_MERGE: Tools for Distributed Explicit State Space
      Generation, March 2006.
[6] G. Gardey, O. H. Roux and O. F. Roux. State space computation and
      analysis of time Petri nets. Theory and Practice of Logic
      Programming (TPLP). Special Issue on Specification Analysis and
      Verification of Reactive Systems, 2005.
[7] Hanna Klaudel and Franck Pommereau. M-nets, a survey. Acta
      Informatica, To appear.
[8] L. Kristensen and L. Petrucci. An approach to distributed state
      space exploration for coloured Petri nets. ICATPN'2004, LNCS
      3099. Springer, 2004
[9] F. Loulergue, F. Gava, and D. Billiet. Bulk Synchronous Parallel
      ML: Modular Implementation and Performance Prediction. In,
      International Conference on Computational Science (ICCS 2005),
      Part II, number 3515 in LNCS, pages 1046-1054. Springer, 2005.
[10] M. Mäkelä. Modular reachability analyzer for high-level
      Petri nets. 5th Workshop on Discrete Event Systems. Kluwer
      Academic Publishers, 2000.
[11] D.Petcu, Parallel explicit-state reachability analysis and state
      space construction, Proceedings of Second International
      Symposium on Parallel and Distributed Computing, ISPDC 2003,
      13-14 October 2003, Ljubljana, IEEE Computer Society Press, Los
      Alamitos
[12] J.-M. Couvreur and D. Poitrenaud. Dépliage pour la
      vérification de propriétés temporelles. In
      Vérification et mise en oeuvre des réseaux de Petri - Tome
      2, chapter 3, pages 127-161. Hermès, 2003.
[13] Model Checking, Edmund M. Clarke, Jr., Orna Grumberg and Doron A.
      Peled, MIT Press, 1999
[14] D. Rémy. Using, Understanding, and Unravellling the OCaml
      Language. In G. Barthe, P. Dyjber, L. Pinto, and J. Saraiva,
      editors, Applied Semantics, number 2395 in LNCS, pages 413-536.
      Springer, 2002.
[15] R. Bisseling, Parallel Scientific Computation. A Structured
      approach using BSP and MPI, Oxford University Presse, 2004.
[16] D. B. Skillicorn, J. M. D. Hill, and W. F. McColl. Questions and
      answers about BSP. Scientific Programming, 6(3): 249-274
[17] Franck Pommereau Versatile Boxes: a Multi-Purpose Algebra of
      High-Level Petri net Proc. of DASD'07, SCS/ACM, 2007
[18] C. Chaouiya. Petri net modelling of biological networks.
      Briefings in Bioinformatics. To appear.