Automatic Test Data Generation Using Genetic Algorithm for Python Programs

In the software industry, testing operation represents a challenging task that is independent of software product quality control. Traditional and common software testing methods are time and resource consuming, especially in manual and human-based methods. In this paper, an automated software test case data generation is proposed based on Genetic Algorithms (GA). Random population generation is used to produce new test data generations based on fitness selection and with different genetic operators by either crossover schemes, mutation schemes, or both of them. The proposed methodology is applied using Python programming language and is tested with four different programs, and by using Distributed Evolutionary Algorithms in Python (DEAP), in addition to unit testing and coverage libraries. The experiments' results prove the advantage of the proposed mechanism over manual testing while results variations are maintained with different crossover and mutation deployments and optimal deployment values are investigated in terms of code coverage percentage.