Model based software testing presentation
Download as PPTX, PDF0 likes608 views
Model Based Testing involves generating test cases from design and analysis models like UML diagrams and finite state machines. As software grows in complexity and size, testing requires more time and effort, so test case generation is automated. UML based testing uses genetic algorithms to generate test data from UML state diagrams before coding begins. The genetic algorithm represents test data as sequences of triggers that fire transitions. It aims to find high quality test data with the best transition coverage through selection, crossover and mutation of sequences over multiple iterations.
Model based software testing presentation
- 1. Model Based Software Testing
- 3. DescriptionModel Based Testing is the generation of test cases and evaluation of test results based on design and analysis models.UML based test case generationFSM based test case generationMarkov chain based test case generation
- 4. ImportanceAs the complexity and size of software grow, what happens ??Time required for testing growsThe effort required for testing also growsWe try to automate test case generationProblem that has received highest attention is test case selection
- 5. UML BASED SOFTWARE TESTINGUsing Genetic Algorithm
- 6. Generating Test Data from software specificationDisadvantage in generating data from source codeDifficulty in extracting behavior of a class from its codeAll different sequences in which messages may be interchanged is very difficult to extract from the codeHaving Test Data before codingUML is one of the software specifications
- 7. Genetic Algorithm for optimizationTest data to be generated is a sequence of triggers.Quality of test dataFinding the best sequence of triggersUML state diagram as the software specification and Genetic Algorithm as the technique for generating test data.The fitness function used is the transition coverage
- 9. Test Data Generation FrameworkObjects should have their status called stateOne object can have one state at a timeUML design uses many diagrams each of which describes a different aspect of the systemState diagrams present the most interesting aspect from testing point of viewConcept of triggers
- 10. GA as a possible solution for searching test dataTest data is a sequence of trigger for executing the state machine diagramThese sequences help testers to know how they should test the systemQuality of test data measured by transition coverage levelThe more transitions test data can fire, the better the quality of data isGenetic Algorithm is selected for searching the best test data
- 11. Genetic Algorithm DesignGA is an iterative solution which works with population of candidatesDuring each iteration selection operator is used to 2 solutions from the current populationSelection being qualified by fitness functionCrossoverMutation