Multi-Objective Test Case Generation for Web Applications with Limited Resources

With the development of web technology, web applications are becoming more complicated and diverse, which makes full testing need lots of resources. Therefore, industries often face how to do effective web application testing with limited time and resources. Evaluating the importance of a web application's page nodes and testing high-importance nodes first is a workable solution. However, the existing methods only consider the content importance on the view of static and local, which lack the consideration of dynamic execution process and the overall topology structure. Web applications usually contain a large number of interaction events with users, which need to dynamically execute server-side logic code to respond to them and update the page nodes. Meanwhile, whether the nodes are in the key position of information transmission in the topology structure of the whole model is also very important for evaluating the importance of web nodes. Thus, in this paper, we propose a novel evaluation method of node comprehensive importance for web applications, which introduces the server-side code coverage to reflect the dynamic information of web applications, the betweenness centrality to consider topology information from a global perspective, and uses the multi-attribute decision-making method to calculate the final comprehensive importance. In addition, this paper implements multi-objective test case generation with GA and NSG A2 algorithms guided by diversity and page importance under the condition of limited resources. This can ensure that the generated test suite has good quality, while also the nodes with higher importance degrees are tested first. In detail, it can cover all important nodes using fewer test cases, with an average reduction of 1.14, compared with the existing methods.