Project-Abstraction:Approximated Inference of Linear Invariants

Team Abstraction

Members

Overall Objectives

Overall Objectives
Highlights of the Year

Scientific Foundations

Abstract Interpretation Theory
Formal Verification by Abstract Interpretation
Advanced Introductions to Abstract Interpretation

Application Domains

Certification of Safety Critical Software
Security Protocols
Abstraction of Biological Cell Signaling Networks

Software

The Astrée Static Analyzer
The Apron Numerical Abstract Domain Library
Translation Validation
ProVerif
CryptoVerif

New Results

Space Software Validation using Abstract Interpretation
Why does Astrée scale up?
Design of the Apron Library
Hierarchy of Semantics by Abstract Interpretation and Formal Proofs
Semantics Design by Abstract Interpretation
Verification of Security Protocols in the Formal Model
Verification of Security Protocols in the Computational Model
Verification of Security Protocols: Formal Model and Computational Model
Analysis of Biological Pathways
Shape analysis of a realistic memory model
Checkmate: a Generic Static Analyzer of Java Multithreaded Programs
Static Analysis via Abstract Interpretation of Multithreaded Programs
SubPolyhedra: A (more) scalable approach to infer linear inequalities
Refining Abstract Interpretation-based Static Analyses with Hints
Approximated Inference of Linear Invariants
An Abstract Domain to Infer Interval Linear Relationships

Contracts and Grants with Industry

ES_PASS Contract
SSVAI Contract
Asbaprod Contract
Contract with CELAR
Survol project
Ascert project
Sardanes project
Astrée exploitation license agreement

Other Grants and Activities

Controvert ANR
FormaCrypt ARA
Thésée ANR
AbstractCell ANR

Dissemination

Interaction with the Scientific Community
Teaching
PhD theses
Participation in Conferences and Seminars
Short-Term Visitors

Bibliography

Major publications
Publications of the year
References in notes

Inria / Raweb 2009
Presentation of the Project Abstraction

Section: New Results

Approximated Inference of Linear Invariants

Participant : Axel Simon.

The Two-Variable-Per-Inequality abstract domain can be used to infer linear invariants of the form ax_i + bx_j $\le$ c where $Im1 ${a,b,c\#8712 \#8469 }$$ and x_i, x_j are two (program) variables. The domain is more efficient than the classic polyhedra domain and more expressive than the Octagon domain. A journal version [18] of the previous conference paper [85] has been accepted pending minor revisions.

Logo Inria