global optimization, differential evolution, individual-dependent strategy, population aaptation

A new variant of individual-dependent differentialevolution (IDE) algorithm is proposed. The original IDE is nhanced by a new mutation strategy accelerating convergence in the last phase of the search. Moreover, the population size is adapted with respect to the diversity of the currentpopulation. The newly proposed IDEbd algorithm is applied to the benchmark suite for CEC 2017 competition on Single Objective Real-Parameter Numerical Optimization. Preliminary experiments showed better performance of IDEbd compared tothe original IDE. The results achieved on the CEC 2017 test suite are also promising, especially in problems of lower dimension.