Abstract
Model-based testing techniques often select test cases according to test goals such as coverage criteria or mutation adequacy.
Complex criteria and large models lead to large test suites, and a test case created for one coverage item usually covers
several other items as well. This can be problematic if testing is expensive and resources are limited. Therefore, test case
generation can be optimized in order to avoid unnecessary test cases and minimize the test generation and execution costs.
Because of this optimization the order in which test goals are selected is expected to have an impact on both the performance
of the test case generation and the size of resulting test suites, although finding the optimal order is not feasible in general.
In this paper we report on experiments to determine the effects of the order in which test goals are selected on performance
and the size of resulting test suites, and evaluate different heuristics to select test goals such that the time required
to generate test suites as well as their size are minimized. The test case generation approach used for experimentation uses
model checkers, and experimentation shows that good results can be achieved with any random ordering, but some improvement
is still possible with simple heuristics.
[download the pdf file] [DOI]