Impact of crossover and mutation on reproduction in evolutionary test model learning

Abstract

Automation in the software test design process has a significant impact on the software testing process and therefore also on the overall software development in the industry. The focus of this paper is on the automation of test case design via model-based testing for automotive embedded software. A method based on an evolutionary algorithm for acquiring the necessary test model automatically from sample test cases and additional sources of information is briefly described. This paper further investigates the impact of reproduction configuration on the evolutionary learning method.