Differential evolution, success-history adaptation, competing strategies, experimental comparison, CEC 2011 real-world test problems.

An enhanced adaptive differential evolution algorithm is described and applied to the CEC 2011 test suite of real-world optimization problems. The new algorithm combines success-history based adaptation known from SHADE and the competition of the various strategies in differential evolution. The comparison of the newly proposed algorithm (called SHADE4) with the algorithm winning in CEC 2011 competition shows that the new algorithm performs signicantly better in 11 out of 22 test problems and worse only in four problems. The performance of SHADE4 was also signicantly better than the original SHADE. The new algorithm is almost control-parameter free, which is helpful to its usage in the solution of the real-world problems.