Automated Test Data Generation for Coupling Based Integration Testing of Object Oriented Programs Using Evolutionary Approaches

Abstract

Software testing is one of the most important phases in development of software. Software testing detects faults in software and ensures quality. Software testing can be performed at unit, integration, or system level. Integration testing tests the interactions of different components, when they are integrated together in specific application, for the smooth functionality of software system. Coupling based testing is an integration testing approach that is based upon coupling relationships that exist among different variables across different call sites in functions. Different types of coupling exist between variables across different call sites. Up until now, test data generation approaches deal only unit level testing. There is no work for test data generation for coupling based integration testing. In this paper, we have proposed a novel approach for automated test data generation for coupling based integration testing of object oriented programs using genetic algorithm. Our approach takes the coupling path as input, containing different sub paths, and generates the test data using genetic algorithm. We have implemented a prototype tool E-Coup in Java and successfully performed different experiments for the generation of test data. In experiments with this tool, our approach has much better results as compared to random testing.