ElitGA : Elitism Based Genetic Algorithm for Evaluation of Mutation Testing on Heterogeneous Dataset

Abstract

Manually generating test cases is a tedious and time-consuming task. Automation testing data production, may help in the creation of a sufficient test suite that meets set objectives. The fault-finding behavior of a test suite determines its quality. For the creation of test data, mutants have been extensively recognized for modelling synthetic faults that act identically to actual ones. The use of search-based strategies to improve the quality of test suites has been widely covered in previous publications. Symmetry, on the other hand, might have a negative influence on the complexities of a search-based technique, whose success is highly dependent on the developing and evaluation of search process. In order to fulfil market expectations for quicker delivery and better-quality software, automation testing has really become critical in the software business. In this work, we proposed a multi mutant evaluation technique using a genetic algorithm. In this work, we carried out a generation of unique test mutants in the first section by using a random population generation algorithm. In the second section vi define a genetic algorithm that performs crossover function, mutation, calculation of fitness and selecting the best jeans according to the percentage of selection. We also define an algorithm for the selection of unique test suites. In the extensive experimental analysis, we evaluate 10 mutants on four different test suites. The proposed genetic algorithm validates all test methods to each test suite and obtains the results whether the mutant has been killed or live. According to this experimental analysis finally, we conclude the effectiveness of the written test suit. The proposed system provides higher efficiency who was the traditional mutation testing evaluation techniques on the heterogeneous datasets.