Maximizing vector distances using differential evolution - Relation to data redundancy

Abstract

This paper studies how redundant data affect maximizing of weighted distances of vectors in a set of vectors. To maximize distances differential evolution is used, because the problem does not have analytical solution and is complex. This paper at first describes suppressing of redundant data mathematically and then it checks this theoretical result in two experiments practically. As a result it was found that both experiments are in correspondence with theory.