Title: Search based test case selection
Shin Yoo

Abstract


A standard approach to the multi-objective problem is to adopt
Pareto-optimality. According to Pareto-optimality, solution A is
better than solution B when A excels B in at least one of the
objectives, while being equal to B in every other
objectives. Therefore it is possible to obtain a group of solutions
that do not dominate each other, instead of a single solution. These
non-dominating solutions are called Pareto-front.

Retrieving a multi-objective Pareto-front from a regression test suite
will provide the tester with multiple alternatives for the procedure
of the testing. Having multiple testing scenarios also guarantees a
certain level of robustness of the testing. Finally, it may be
possible to get some meaningful insight into the properties of the
given regression test suite and its target program by observing the
Pareto-front itself.