Formal verification in the interactive and intelligent systems
No Thumbnail Available
Date
2024
Authors
Journal Title
Journal ISSN
Volume Title
Publisher
Université Badji Mokhtar Annaba
Abstract
Computer modeling and simulation of human reasoning, decisions and behavior is an important goal of Artificial Intelligence. Much progress has been made in this direction since the emergence of this research problem and several researchers have proposed works using modal logic to represent human reasoning. One of the best known is the BDI model (for Beliefs, Desire, Intentions) proposed by Bratman in the 1980s. In this model, the designer represents the beliefs of individuals and the system builds plans from the goals.
In this thesis, there is interest in the use of this reasoning model in the context of modeling and automatic verification of systems. Modeling and verification of systems consists in describing the operation of a system, using logical formulas that can be interpreted by a computer, in order to verify certain properties. This model checking technique has given good results for several years in critical systems where the slightest error can have serious consequences. However, these realized tools do not take into account human factors such as beliefs, emotions or personality. However, these can be the source of serious failures in the systems in which the human being is part. One of the most famous examples is the Mont Saint-Odile crash where an erroneous belief about the state of the system led to the death of 97 people despite two very well-trained pilots and an aircraft in perfect working order. Our objective is to propose a verification model that integrates these human factors to better prevent this type of failure.
Description
Keywords
BDI; formal verification; model checking; EBDI; NUSM