Optimization of Test Data for Basis Path Testing using Artificial Intelligence Techniques

Abstract

Software testing is a process carried out with the intent of finding errors. This helps in analyzing the stability and quality of a software. Stability and quality can be achieved by suitable test data. Test data can be generated either manually or by automated process. Manual generation of test data is a difficult task. It involves lot of effort due to presence of huge number of predicate nodes in a module. In this report, an automated process is proposed for test data generation in traditional methodology for the automatically constructed control flow graph. Code Coverage is a measure used in software testing process and is one of the key indicators of software quality. It helps the tester in evaluating the effectiveness of testing. It is achieved by automatically generating test data for various functions. Code coverage is not a method or a test; it is a measure which helps in improving software reliability. Effort has been made to gather code coverage information either by source code or by the requirements specified by the customer. But less attention has been paid to achieve better coverage. This report also emphasizes on code coverage, achieved through the test data generated, using some soft computing techniques. Here, three soft computing techniques such as Genetic algorithm, Particle swarm optimization and the Clonal selection algorithm techniques have been deployed for automatic test data generation. This test data was in turn used for code coverage analysis. Experimental results show that the test data generated using Clonal selection algorithm was much more effective in achieving better code coverage over Genetic algorithm and Particle swarm optimization.