2019-12-05 • salle des thèses, Faculté des sciences Jean Perrin

Résumé

This work presents four of the most important research results obtained after my Ph.D., i.e., from 2008 until 2019. After a brief state of the art, a formalism aiming at modelling responsibility in multi-agent environments is put forward. This formalism, called CEDL, is a modal logic that contains epistemic and dynamic operators as well as some other operators defined as abbreviations, such as obligation, agents abilities and agents knowing how abilities. The second result is an evolution of CEDL, called ATDEL. This logic aims at modelling agents abilities and knowledge trough time. It permits system specifications that are much smaller than CEDL. In addition, apart from an axiomatisation, algorithms for model checking and satisfiability checking of formulas are also provided. The third result is an adaptation of AGM belief revision theory to multi-agent scenarios. Multi-agent versions of expansion and revision postulates are proposed. The last result presented in this work is about modal logic automated reasoning. Methods for satisfiability checking of formulas in K and KT5 are proposed. Practical experiments show that our technique outperforms alternative approaches.

Rapporteurs

  • Anthony Hunter, professeur à l’University College London, Royaume-Uni
  • Odile Papini, professeur à l’Université Aix-Marseille, France
  • Umberto Straccia, directeur de recherche au CNR, Italie

Examinateurs

  • Salem Benferhat, professeur à l’Université d’Artois, France
  • Philippe Besnard, directeur de recherche au CNRS, France
  • Sébastien Konieczny, directeur de recherche au CNRS, France