Using Formal Methods for Test Case Generation According to Transition-Based Coverage Criteria

Abstract

- Formal methods play an important role in increasing the quality, reliability, robustness and effectiveness of software. Also the uses of formal methods, especially in safety-critical systems, help in the early detection of software errors and failures which will reduce the cost and effort involved in software testing. The aim of this paper is to prove the role and the effectiveness of formal specification for the cruise control system (CCS) as a case study. A CCS formal model is built using Perfect formal specification language, and its correctness is validated using the Perfect Developer toolset. We develop a software testing tool in order to generate test cases using three different algorithms. These test cases are evaluated to improve their coverage and effectiveness. The results show that random test case generation with full restriction algorithm is the best in its coverage results; the average of the path coverage is 77.78% and the average of the state coverage is 100%. Finally, our experimental results show that Perfect formal specification language is appropriate to specify CCS which is one of the most safety-critical software systems, so the process of detecting all future possible cases becomes easier.