GA-based Automatic Test Data Generation for UML State Diagrams with Parallel Paths

C. Doungsa-ard, K. Dahal, A. Hossain, T. Suwannasart

    Research output: Chapter in Book/Report/Conference proceedingChapterpeer-review

    23 Citations (Scopus)

    Abstract

    Automatic test data generation from a software specification prepares test cases for software developers before their code development phase. Having test cases before coding, helps the developers to control their code to conform to the specification. In state-based specifications, paths from the initial state to the final state may be varied, this is called parallel paths. In order to generate test data which cover all behaviors in the specification, test data for each parallel path should be generated. This paper proposes an enhanced genetic algorithm(GA)-based approach to resolve the parallel paths from the UML state machine diagram. The proposed approach is improved from our previous study which uses a GA based test data generation method for only one path. The approach identifies parallel paths to cover all transitions. GA evolves a number of suitable test data sets - one test data set for each parallel path. The best test data for each parallel path are picked and used for calculating the overall coverage test data. The experimental results show improved coverage results with the enhanced approach for a number of case studies with parallel paths.
    Original languageEnglish
    Title of host publicationADVANCED DESIGN AND MANUFACTURE TO GAIN A COMPETITIVE EDGE
    EditorsX.T. Yan, C. Jiang, B. Eynard
    PublisherSpringer-Verlag
    Pages147-156
    ISBN (Print)978-1-84800-240-1
    DOIs
    Publication statusPublished - 2008

    Keywords

    • Test data generation
    • UML state machine diagram
    • Genetic algorithm

    Fingerprint

    Dive into the research topics of 'GA-based Automatic Test Data Generation for UML State Diagrams with Parallel Paths'. Together they form a unique fingerprint.

    Cite this