A Glance on Performance of Fitness Functions Toward Evolutionary Algorithms in Mutation Testing

Abstract

Nowadays, Internet and web applications have influenced on different aspects of human life. Therefore there are always some needs to different software platforms for implementation of electronic commerce or electronic governance. Hence a great market is now devoted to software production in various platforms. Regarding such market demand, producing high-quality softwares with reliability, safety and availability services are considered as an important issue. To be more specific all software companies use software testing concepts as an independent process in software development cycle. There are various methods for software testing, but mutation testing is one of the most powerful tools. In mutation testing, high-quality test-case generation plays a key role and it has a direct relation with quality of software testing. There are different techniques for test-case generation where evolutionary algorithms are among the most common ones. Since each evolutionary algorithm needs an appropriate fitness function which is dependent on target problem, it is very important to know that for each evolutionary algorithm which fitness function generates better test cases. The main goal of this paper is to answer this question and a treatment of five evolutionary algorithms regarding four different fitness functions are classified in this work.