A new variant of individual-dependent differential evolution (IDE) algorithm is proposed. The original IDE is enhanced 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 current population. 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 to the original IDE. The results achieved on the CEC 2017 test suite are also promising, especially in problems of lower dimension.