A Survey of Methods for Automated Algorithm Configuration

Algorithm configuration (AC) is concerned with the automated search of most suitable parameter a parametrized algorithm. There currently wide variety AC problem variants and methods proposed in literature. Existing reviews do not take into account all derivatives problem, nor they offer complete classification scheme. To this end, we introduce taxonomies to describe features methods, respectively. We review existing literature within lens our taxonomies, outline relevant design choices approaches, contrast against each other, state industry. Finally, provides researchers practitioners look at future research directions field AC.