Data Flow Based Automated Integration Testing Framework for Object Oriented Programs Using Evolutionary Approach"

Abstract

Software is widely used technology of this era. Applications are being developed, solving tasks ranging from simple calculations to complex banking transactions and so on. Use of software in critical systems leaves no space for even simple errors which can lead to sever nancial loss or threat to human life. Software must be well tested to verify its functioning. Manual testing is tedious and time consuming job. Object oriented programs use classes and their interactions to perform tasks. Integration testing requires to test the interfaces. Path explosion make it worst to test the integration of classes. Small work done in integrated test case generation of OO programs. Coupling criteria is used for generation of automated test cases in proposed approach. Search space is reduced by selecting only coupling methods that are directly involved in integration. Objects must be in proper state before testing process started. Further def-use analysis helps in achieving the desired object states for proper interfaces testing, representing methods in intermediate tree. Method sequences are generated considering data ow of involved state variables. Test can't be accomplished without proper input data. We propose a tness function for test data generation considering coupling path coverage. Genetic algorithm is used to optimize the solution based on proposed tness function. Eleven randomly selected project are used from SF100 (Software Testing Benchmark) to show the strength of proposed approach. Results showed relatively high coverage as compared to random testing.