Search

Differential Evolution (DE) is one of the most popular and powerful branches of evolutionary algorithm family. However, even many state-of-the-art DE-based variants still exist weakness such as improper parameter adaptation and population stagnation during the later stage of evolution. To mitigate these deficiencies, differential evolution with fitness-difference based parameter control and hypervolume diversity indicator (FDHD-DE) is proposed in this paper. Firstly, a semi-adaptive adaptation scheme for control parameters is proposed, in which the generation of scale factor and crossover rate is modified by dividing into two stages, thus enhancing the efficiency of parameter adaptation. Secondly, a novel fitness-based weighting strategy is proposed to improve the performance of existing success history-based adaptation by employing a novel approach of utilizing fitness information. Finally, a hypervolume-based diversity indicator and corresponding dimension exchange strategy are proposed to alleviate the problem of population stagnation. The performance of FDHD-DE is verified on the 88 benchmark functions from Congress on Evolutionary Computation (CEC) 2013, CEC 2014, and CEC 2017 test suites on 10D, 30D and 50D and a real-world application. The experiment results are compared with several state-of-art DE variants, and the results show that FDHD-DE has better performance, both in terms of solution accuracy and convergence speed. [ABSTRACT FROM AUTHOR]

Differential Evolution (DE) is a potent population-based global optimization algorithm which has already proven efficient for optimization demands in engineering applications. However, even the state-of-the-art DE variants still suffer premature convergence and lack of diversity. In order to overcome the above mentioned weaknesses, this paper proposes a brand-new DE variant, namely zDE algorithm, for single-objective numerical optimization. The main contributions can be summarized as follows: First, an improved wavelet basis function is incorporated into the generation of the scale factor F and a Minkowski Distance based adaptation scheme is proposed for the adaptation of it. Second, a new trial vector generation strategy is first proposed as a supplementary of the existing strategies, and t-distribution is incorporated into this strategy. Third, a novel diversity enhancement technique is firstly proposed by changing the dimensional parameters of the individuals in the population. The zDE algorithm is validated under 88 benchmarks from the CEC2013, CEC2014, and CEC2017 test suites for real-parameter single-objective optimization, and the results show the superiority of our algorithm in comparison with the recent state-of-the-art DE variants. • First, novel wavelet basis function is employed in the generation of scale factor F. • Second, Minkowski distance based adaptation is proposed for parameter control. • Third, a novel mutation strategy with t-distribution based perturbation is developed. • Fourth, a new crossover method is proposed to maintain population diversity. • Fifth, novel stagnation indicator and dimensional interchange are proposed. [ABSTRACT FROM AUTHOR]