PERFORMANCE ANALYSIS OF PRIORITIZED TEST SUITES BASED ON FAULT DETECTION
Usha Badhera, Annu Maheshwari, M. Tech Scholar · 2014
Specification based testing identifies test cases from software requirement specifications. This leads to better quality software which reduces effort and cost. Test cases generated for Boolean specifications have been widely used to specify requirements of safety critical softwares, avionics, medical and other control software. Various Boolean specification techniques have been proposed among them MC/DC and MUMCUT techniques are the popular testing techniques. Compliance of the MC/DC criterion has been mandated by Federal Aviation Administration for the approval of airborne software. According to Kaminski (2010) the Federal Aviation Administration requires the Minimal-MUMCUT criterion instead of MC/DC for Irredundant Disjunctive Normal Form (IDNF). The Minimal-MUMCUT criterion provides better logic fault detection .In this paper performance analysis of proposed prioritized test suite generated from Minimal MUMCUT has been done. The APFD of prioritized test suite is computed and compared with other possible prioritized test suites. Minimal MUMCUT identified testing techniques for Boolean specification for which various 2 n distinct Boolean functions with n variables can be formed. To distinguish one from all others using exhaustive testing, it would require 2 n distinct test cases. Test cases generated by Minimal MUMCUT are less than the test cases generated by MUMCUT Strategy. The proposed approach for prioritization of test cases generated by Minimal MUMCUT yields higher APFD and hence early detection of faults.