Abstract
Formal verification aims to improve the quality of software by detecting errors before they do harm. At the basis of formal verification is the logical notion of correctness, which purports to capture whether or not a program behaves as desired. We suggest that the boolean partition of software into correct and incorrect programs falls short of the practical need to assess the behavior of software in a more nuanced fashion against multiple criteria. We therefore propose to introduce quantitative fitness measures for programs, specifically for measuring the function, performance, and robustness of reactive programs such as concurrent processes.This article describes the goals of the ERC Advanced Investigator Project QUAREM. The project aims to build and evaluate a theory of quantitative fitness measures for reactive models. Such a theory must strive to obtain quantitative generalizations of the paradigms that have been success stories in qualitative reactive modeling, such as compositionality, property-preserving abstraction and abstraction refinement, model checking, and synthesis. The theory will be evaluated not only in the context of software and hardware engineering, but also in the context of systems biology. In particular, we will use the quantitative reactive models and fitness measures developed in this project for testing hypotheses about the mechanisms behind data from biological experiments.
Highlights
This article describes the goals of the European Research Council (ERC) Advanced Investigator Project QUAREM
This article describes the goals of the ERC Advanced Investigator Project QUAREM
The boolean framework is based on binary satisfaction relations between reactive systems and behavioral requirements, and on binary refinement relations between reactive systems
Summary
The project aims at rebuilding a central part of the formal foundation of computing by replacing the classical, boolean notion of program correctness with a new, quantitative measure of program fitness. One program is often preferred over another, even if both are technically correct (for example, one may be more robust against faulty inputs than the other), or if both are technically incorrect (one may misbehave less often, or less severely, than the other). Such behavioral preferences can be formalized by quantitative measures of fitness between programs and specifications. In biology the use of computational models for testing mechanistic hypotheses has been hampered by the lack of quantitative measures of fitness between models and experimental data
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
