Novel Approach to Improve Concolic Testing

摘要

Concolic Testing is the combination of symbolic as well as concrete execution. It considers program variables as symbolic variables along with concrete execution path. Branch Coverage belongs to white box testing. Its objective is to demonstrate that all conditions present in a predicate can impact the estimation of predicates in a particular manner. In the area of aerospace and safety critical domains, software quality assurance is strict to specified rules and regulations that maintained in DO-178 standard. To resolve such issues, Concolic testing generates automated test cases to attain high branch and MC/DC coverage in an automated technique based on the branch coverage. Here, we consider Java programs for achieving a high percentage. We are taking Java program as input that is named as J. This Java program is supplying to JPCT (Java program code transformer) which gives the output program called transformed program J which is feed as input to JCUTE (Java Concolic tester tool) to generate test cases as well as to achieve more branch coverage percentage. Our study here is based on two steps, and one is taking Java program without the help of JPCT, and other is with the help of JPCT. The percentage coverage of branch is more in the transformed program without affecting the output with automatically adding some extra statements to the input. This resolves some of the bottleneck issues of the previous Concolic testing methods. We are getting the result experimentally show that our method achieves an average 17.21 percentage more branch coverage.