الفهرس 
INTRODUCTIONOnly 14 pages are availabe for public view
 |  | from | 132 |  |  |
| | | from | 132 | | |
Abstract
ABSTRACT
Software testing is accounted to be an important phase in software development life cycle in terms of cost and manpower. It is considered the key to success of any software. Consequently, many studies have been conducted to minimize the associated cost and human effort to fix bugs and errors, and to improve the quality of the testing process by generating test cases at early stages. There exist many software development models, as waterfall model, agile model, etc. Test cases can be generated at different phases (analysis phase, design phase and after development phase). Though, test cases generation at early stages is more effective rather than that after development, where time and effort used for finding and fixing errors and bugs are less than that after development. At a later stage, fixing errors results in enormous code correction, consuming a lot of time and effort.
Requirements-based testing is a testing approach in which test cases are derived from the requirements. Requirements represent the initial phase in software developments life cycle. Requirements are considered the basis of any software project. Therefore, any ambiguity in natural language requirements leads to major errors in the coming phases. Moreover, poorly defined requirements may cause software project failure. Model-based testing (MBT) is the automatic generation of software test procedures, using models (UML diagrams) of system requirements and behavior. Whilst Search Based Software Testing (SBST) is a branch of Search Based Software Engineering (SBSE), in which optimization algorithms are used to automate the search for test data that maximizes the achievement of test goals, while minimizing
ii
testing costs. However, most of related studies considered only one type of behavioral diagrams with a lot of human intervention.
In this thesis, we study the different paradigms of testing techniques for generating test cases, where we investigate their coverage and associated capabilities. We then propose a consolidated model for a generic automated test cases generation. The model consists of two main parts to handle the UML diagrams, the software requirements specification (SRS) documents generated from the waterfall development model and the user stories generated from the agile development model, which are considered as the main deliverables of the analysis and design phases in the software development life cycle.
In the first part (UML Test Cases Generation Part), an optimized automated approach for generic and dynamic test cases generation is proposed. It is generic and dynamic as it can be applied on different types of behavioral diagrams (i.e. activity diagram, state diagram, uses case diagram, etc) for multi-disciplinary domains. While automation is considered to generate test cases with minimum human intervention, which will consequently help to minimize total cost. An optimization technique is applied to optimize the generated test cases to ensure the quality of results. The proposed approach merges model-based testing with search-based testing to automatically generate test cases from different behavioral diagrams, i.e. use case, activity, etc. Whereas in the second part (Requirements Test Cases Generation), we propose another novel automated approach to generate test cases from requirements. Requirements can be gathered from different models either waterfall model (functional and non-functional) or agile model. SRS documents, non-functional requirements and user stories are parsed to generate test
iii
cases. The proposed consolidated model uses text mining and symbolic execution methodology for test data generation and validation, where a knowledge base is developed for multi-disciplinary domains. The proposed model is a time saving model where it will take about 70 minutes (4200000 milliseconds) to perform all the proposed steps manually for UML Test Case Generation Phase. While it takes about 0.204 minutes (12269 milliseconds) when using our automated system to perform the same steps which ensure time and cost saving.