Modified Genetic Algorithm for Regression Testing

Abstract

Software Testing is an approach where different errors and bugs in the software are identified. In this thesis, we have developed the approach to generate test cases automatically from some initial random test cases using Evolutionary Algorithms (EA). As evolutionary algorithms are known to get the optimized results, so we are using genetic algorithm to get the optimal results. To test software we need the test cases. One of the most important activities in software maintenance is Regression Testing. The reexecution of all test cases during the regression testing is costly and time consuming. And even though several of the code based proposed techniques by researchers address procedural programs. In our research work we proposed a regression test case selection which optimizes the selected test case using Genetic Algorithm. We are executing genetic algorithm upon different crossover rates (CR) and analyzing the results on number of iterations. The test cases are automatically generated through path crawler tool. We have taken 100% path coverage of the given source code. The effectiveness of the approach was evaluated calculating Average Percentage of Modified Genetic Algorithm (MGA) over Simple Genetic Algorithm (SGA). Our Proposed Approach (PA) provides considerably better results in term of average percentage.