Concurrent fault localization using ANN

The software is becoming more capable of providing better solutions to our day-to-day activities. In order to increase performance, concurrent programs are always preferred. But the concurrency produces obstacles to different phases of software development, including testing and debugging. Finding faults in a concurrent program is always a challenging task due to the presence of many threads overlapping each other, problems in sharing memories, etc. In this paper, we have proposed a Back propagation neural network (BPNN) to generate ranks for each class of a given program. These ranks indicate the probability of a fault being present in each class. The model is trained using test case coverage data, and it is tested using virtual test cases. In all three case studies, the fault localization technique assigned the highest rank to the classes where the actual faults were implanted. By using BPNN and training the model with test case coverage data, the technique shows promising results in identifying the classes where faults are likely to occur. This can greatly aid in the testing and debugging processes of concurrent programs, improving their overall performance.