Test Scenarios Generation and Optimization of Object-Oriented Models Using Meta-Heuristic Algorithms

Abstract

Software testing offers re-utilization of the model for the operation of validation and this leverages the test case generation growth. It is a very crucial and complicated action in software establishment as it is very concise with software standards. The testing process is comprised of three main parts, such as test scenarios generation, test implementation, and test assessment. The process of test case generation serves a significant part in all three circumstances. The major components of the test case are input to the module, the state of the module and the targeted outcome. If the test case finds numerous faults with very few test cases, then it is stated as better coverage. During the software development mechanism, testing can be executed at any time and anywhere, but the testing is executed after the needs are described and the coding mechanism is completed. Automatic testing is utilized to occupy most resources, like cost, effort, and time. Usually, behavior illustration and Unified Modeling Language (UML) structural diagrams are used by researchers for test case generation at the early phase of evolution. This mechanism efficiently ensures the durability of the system with the assistance of improved test coverage. Software testing utilizing an object-oriented model is a challenging task among the research community in the modern period. In order to enhance the standard of the software, automation of testing has become a crucial part. Hence, this research provides a unified solution for best test case generation in an object-oriented model. In the first contribution, the method proposes an approach to generate the test scenarios from the integrated models of sequence and state machine diagrams with the help of a case study. This method is systematic and highly logical. The developed approach is very efficient in dealing with errors in the loop and inaccurate message responses. In our second contribution, we have proposed an algorithm to optimize the generated test sequences from UML behavioral diagrams. The sequences that enclose all the test probabilities are chosen by exploiting developed Fractional-SMO, which is newly devised by the amalgamation of Fractional calculus with SMO. Therefore, suitable test cases are selected depending on the optimization that utilizes the factors such as coverage and fault. Finally, in the third contribution, we proposed a hybrid approach called Spider Monkey Particle Swarm Optimization (SMPSO) to optimize the produced test cases from the developed models. Accordingly, the proposed algorithm efficiently produces the best test cases from UML by means of the framing of a control flow graph. However, the proposed algorithm attained a maximum coverage of 85% and is capable of generating maximum test scenarios.