Efficient evolutionary optimization algorithm : filtered differential evolution

Abstract

Solving many real-life engineering problems requires often global and efficient (in terms of objective function evaluations) treatment, because function values involved are produced via time consuming simulations. In this study, we consider optimization problems of this type by discussing some drawbacks of the current surrogate assisted methods and then introduce a new population based optimization algorithm, which borrows features of the well-known Differential Evolution algorithm, but improves its efficiency by filtering away ineffective trial points.