Hochschulschrift
Automated verification of system requirements and software specifications
Zusammenfassung: Many small to medium-sized enterprises that specialise in electrical or communications engineering are challenged by the increasing importance of software in their – often safety-critical – products. For this reason they have a strong interest in subcontracting competent partners for software development tasks. Yet, many of those enterprises refrain from subcontracting because they assessthe associated risks as too high. This thesis presents a set of methods and tools based on formal methods that are useful in reducing those associated risks. We are concerned with software-defined systems, i.e., systems whose behaviour is solely controlled by a program and the input the program receives from the environment. We especially focus on systems whose controlling programaccesses its hardware via memory-mapped I/O.We present a method for validating requirements between stakeholders with different levels of expertise. Based on validated requirements, we present the IRS -method that reduces the problem of deciding whether a software-defined system satisfies the requirements to the problem of whether the controlling program satisfies its specification. We show the feasibility of the IRS -method through an industrial case-study. The case-study uses a class of state-of-the-art tools for proving program correctness (deductive verifiers).
- Standort
-
Deutsche Nationalbibliothek Frankfurt am Main
- Umfang
-
Online-Ressource
- Sprache
-
Englisch
- Anmerkungen
-
Albert-Ludwigs-Universität Freiburg, Dissertation, 2016
- Klassifikation
-
Informatik
- Schlagwort
-
Software Engineering
Formale Methode
Model Checking
Requirements engineering
Statische Analyse
Cyber-physisches System
- Ereignis
-
Veröffentlichung
- (wo)
-
Freiburg
- (wer)
-
Universität
- (wann)
-
2016
- Urheber
- Beteiligte Personen und Organisationen
- DOI
-
10.6094/UNIFR/11159
- URN
-
urn:nbn:de:bsz:25-freidok-111593
- Rechteinformation
-
Der Zugriff auf das Objekt ist unbeschränkt möglich.
- Letzte Aktualisierung
-
25.03.2025, 13:42 MEZ
Datenpartner
Deutsche Nationalbibliothek. Bei Fragen zum Objekt wenden Sie sich bitte an den Datenpartner.
Objekttyp
- Hochschulschrift
Beteiligte
- Dietsch, Daniel
- Podelski, Andreas
- Universität
Entstanden
- 2016