Automatic Derivation of Search Objectives for Test-Based Genetic Programming

Abstract. In genetic programming (GP), programs are usually evalu-ated by applying them to tests, and fitness function indicates only how many of them have been passed. We posit that scrutinizing the outcomes of programs ’ interactions with individual tests may help making program synthesis more effective. To this aim, we propose DOC, a method that autonomously derives new search objectives by clustering the outcomes of interactions between programs in the population and the tests. The derived objectives are subsequently used to drive the selection process in a single- or multiobjective fashion. An extensive experimental assess-ment on 15 discrete program synthesis tasks representing two domains shows that DOC significantly outperforms conventional GP and implicit fitness sharing