Intact Analysis With Compiler And Security Editor
We argue that writing a first set of test cases is easy, and most developers do such basic testing. Our experiments showed that such test cases easily reach 60 % of test quality (see ). Improving test quality implies a particular and specific supplementary testing effort. In this section we investigate the use of genetic algorithms as a pragmatic way to automatically improve the basic test cases set in order to reach a better test quality level with limited effort. Indeed, the basic test cases set carries information that can be optimized to create better test cases, by some cross-checking and “mutation” of the test cases themselves. So, at the beginning we have a population of mutants programs to be killed and a test cases pool. We randomly combine those test cases (or “gene pool”) to build an initial population of test cases seen as predators of the mutant population. From this initial population, we apply a genetic algorithm to improve its ability to kill mutants programs. We propose a new AI algorithm that fits better to the test optimization problem we called bacteriological algorithm (BA): Bas behave better that GAs for this problem. However, between GAs and BAs, a family of intermediate algorithms exists: we explore the whole spectrum of these intermediate algorithms to determine whether an algorithm exists that would be more efficient than BAs.: the approaches are compared on a .Net system.