Code Coverage Test Case Prioritization for Regression Testing Using Non-Deterministic Approach

Doctoral Thesis / Dissertation from the year 2025 in the subject Computer Science - Software, grade: 1, course: CSE, language: English, abstract: Effective test case selection method is important, since it is helpful in scheduling test case groups in a specific order. This method can maximize objective functions of several tests, including maximum code coverage, maximum faults at an early stage, and less test case execution time. The vital part of maintenance is the software development cycle which help to satisfy the quality constraints is regression testing. Regression testing is more time-consuming with respect to executing all available test cases belonging to a specific test suite. This implies that the total cost is considerably increased. In this framework, a test case selects the minimum number of collections that belongs to the test suite, which is capable of identifying the most faults in the minimal time frame. Software testing consists of different forms of objective functions, namely, the detection rate of faults, which measures the time frame required to detect faults as the testing process progresses. When regression testing takes place, the enhanced rate of detecting faults can provide feedback quickly for debuggers so they can start work as soon as possible. The result is the need for a new approach to test algorithms for evaluating the methods which were not invoked. A large portion of research work is connected to the issues mentioned above but it is not possible to find a comprehensive solution for them. This research presents a new framework for the resolution of such problems. Two phases are included in the proposed work including selection of a crossover algorithm, generation of optimal test cases, and testing return type nul functions and to invoke uninvoked methods through the use of NFA (Non-Deterministic Finite Automata). In the final section of the research, our proposed framework is depicted experimentally, which is efficient as well a