An Artificial Immune System for Black Box Test Case Selection.

Testing is a crucial part of the development of a new product. For software validation a transformation from manual to automated tests can be observed which enables companies to implement large numbers of test cases. However, during testing situations may occur where it is not feasible to run all tests due to time constraints. Hence a set of critical test cases must be compiled which usually fulfills several criteria. Within this work we focus on criteria that are feasible for black box testing such as system tests. We adapt an existing artificial immune system for our use case and evaluate our method in a series of experiments using industrial datasets. We compare our approach with several other test selection methods where our algorithm shows superior performance.