Generating Linear Temporal Logics Based on Property Specification Templates

Weibin Luo, Hironori Washizaki*, Yoshiaki Fukazawa

*Corresponding author for this work

Research output: Chapter in Book/Report/Conference proceedingChapter

Abstract

Temporal logics are widely used in software verification such as model checking. However, creating temporal logics such as linear temporal logics (LTLs) based on property specifications written in a natural language is difficult due to practitioners’ unfamiliarity with property specifications and notations of temporal logics. Although property specification patterns have been introduced to help write correct temporal logics, creating temporal logics using property specification patterns requires an understanding of the pattern system. Since some patterns are difficult to understand, especially for beginners, and the final temporal logics are usually complicated, creating temporal logics using pattern systems is time consuming and error-prone. Here, we introduce a method to create LTLs based on property specification patterns. We experimentally compare the required time and accuracy of our approach to those using property specification patterns. Our approach can improve the creation of LTLs in terms of speed and accuracy. Although our experiment is implemented in Japanese, the results should be applicable to other languages such as English. We also provide a visualization scheme so that practitioners can understand the generated LTLs and confirm that they are correct.

Original languageEnglish
Title of host publicationStudies in Computational Intelligence
PublisherSpringer Verlag
Pages1-15
Number of pages15
DOIs
Publication statusPublished - 2020

Publication series

NameStudies in Computational Intelligence
Volume850
ISSN (Print)1860-949X
ISSN (Electronic)1860-9503

ASJC Scopus subject areas

  • Artificial Intelligence

Fingerprint

Dive into the research topics of 'Generating Linear Temporal Logics Based on Property Specification Templates'. Together they form a unique fingerprint.

Cite this