EARLY IDENTIFICATION OF SOFTWARE DEFECTS USING OCL PREDICATES TO IMPROVE SOFTWARE QUALITY

Abstract

Formal specification-based testing has been used widely to assess potential faults or prove their absence in a given system at the earliest. This research work has proposed an automated fault-based testing framework to test the specification of the system using Object Constraint Language (OCL). Accordingly, the possible faults in OCL specification of the system has been anticipated by mutating its method based on OCL predicate-based fault classes. Then, test cases are generated using Genetic Algorithm with simulated annealing technique. In this paper, a novel OCL-predicate based fitness function is defined to evaluate the generated test data. Finally, this paper presents the experimental results, which indicate that the proposed methodology provides more test coverage with the reduced test suite and test run. This results in cost-effective software development so as to improve software quality.

Keywords

• Specification-based testing
• OCL
• Genetic Algorithm
• Fitness function