One of the major research in Computer Science and specifically in Artificial Intelligence
concerns the knowledge representation and the reasoning from such knowledge. A classical
approach is based on the concepts of mathematical logic. The problem is that in computer
science practical results are expected: the deduction must be programmed and produce these
results in a reasonable time. It is in this very general context that my research is done.
In fact, many practical problems can be represented simply in propositional logic. But
the fact remains that the problem SAT (decision problem of the satisfiability a set of
constraints built on a finite set of propositional symbols and expressed in the form clausale)
is NP-complete. At present, the deduction in propositional calculus can not exhibited a
general and efficient algorithm in all cases. Until the P = NP conjecture has not been proved
or disproved, we can not know if a algorithm is able to solve efficiently (ie such that the
time complexity at the worst case is polynomial) this problem as well as those problems which
are naturally related it (automatic proof, logic programming, deductive databases, VLSI, etc).
The main goal of my research is focused on the expansion of the class of treatable problems in
practice by the way of new technics, new heuristics or new algorithms increasing the practical
efficiency of treatments, or by the discovery of new restrictions for wich it is possible to
ensure a resolution in a polynomial time.
My research topics are:
- Knowledge representation and reasoning
- Practical resolution of NP-Hard problems
- Propositional logic: models and algorithms
- Satifiability problem (SAT)
- Quantified Boolean Formulas (QBF)
- Constraint Satifaction Problems (CSP)
- Compilation of knowledge bases
- Localization and treatment of inconsistencies
- Non-monotonic logic
- Cooperative work and fusion of knowledge bases
- Validation of knowledge bases systems
My research is done as part of the axis
"Algorithms for inference and decision process"
of the computer science research center of LENS (CRIL).