Fault Aware Test Case Prioritization in Regression Testing using Genetic Algorithm

Abstract

In software testing practice, regression testing is significant type of testing, which is responsible for stability, quality, reliability and functionality of existing application even after doing the modifications in the application. Test case prioritization (TCP) is one of the proved effective techniques of regression testing that improves defect detection rate by scheduling execution of the test cases (TCs). Scheduling the execution of all TCs is very complex and time consuming task and thus needs to introduce the use of optimization algorithms. This paper implements a genetic optimization algorithm (GA) to improve the TCP technique by ordering the TCs with goal of maximum fault detection by minimal execution of TCs. The effectiveness of implemented GA optimization technique is measured using average percentage of fault detection (APFD) metric. We analyzed implemented GA approach to examine its effect on outcome by changing its vital parameters such as crossover, mutation and convergence criteria with the aim of increasing rate of fault detection. This experiment is evaluated on public dataset with more than 1000 TCs. We tend to compare our work with random search prioritization and hill climbing optimization algorithms. This carried out experimental outcome clearly depict that GA outperforms better than compared algorithms in solving TCP problem by improving the performance of regression testing.