|
Hamoudi Kalla
Assistant Professor of Computer Science University
Habilitation
|
Background
Born in 1977, I am a Chaoui originally from Menaa - a small nice city of
Algeria. I am graduated in Computer Science Engineer since October
2000 at the University of Batna. I received my Ph.D from Institut
National Polytechnique de Grenoble.
Research/Academic Interests
My research interests include, but are certainly not limited to:
- Real-Time Distributed Embedded Systems
- Dependability : Reliability and Fault-Tolerance
- Real-Time Scheduling Algorithms
- Multi-Criteria Scheduling Algorithms
- Verification and Validation of Embedded Systems based Intellectual Property (IP) Components
Since 2001, I am interested in developing new dependability techniques
for distributed real-time embedded systems. These systems are being
increasingly used in critical real-time applications, such as
avionics, air traffic control, autopilot systems, and nuclear plant
control, in which the consequences of missing a tasks deadline may
cause catastrophic loss of money, time, or even human life. My goal is
to propose a solution to automatically produce a fault tolerant and a
reliable code of a given algorithm onto a given distributed
architecture, according to reliability and real-time constraints. As
a result of the research collaboration with
Alain
Girault (scientific leader of the POP ART team)
and Yves
Sorel (scientific leader of the Aoste team ), we
have integrated into SynDEx tool new dependability techniques for hardware
faults. I am interested also with Alain Girault and
Denis Trystram in developing new bi-criteria Scheduling problems.
Since 2005, I am interested with
Jean-Pierre Talpin in developing new techniques for functional and
compositional correctness of Intellectual Property Components. Common
system level capture in software programming languages such as C/C++
allow for a comfortable design entry and simulation, but mere
simulation is not enough to ensure proper design
integration. Validating that reused components are properly connected
to each other and function correctly has become a major issue for such
designs and requires the use of formal methods. we propose an approach
in which we automatically translate C/C++ programs into the
synchronous formalism SIGNAL, hence enabling the application of formal
methods without having to deal with the complex and error prone task
to build formal models by hand. The main benefit of considering the
model of SIGNAL for C/C++ languages lies in the formal nature of the
synchronous language SIGNAL, on which verification and optimization
techniques can be performed.
| |