If an appropriate fitness function can be defined for the test aim, and some search algorithm has been selected, then tests proceed as follows: an initial set of individuals (i.e., test data) is generated, usually at random; for each test data set the SUT is executed and monitored and the fitness value is determined; next, test data with high fitness values are selected with a higher probability than those with a lower value and are subjected to combination and mutation processes to generate new offspring test data.