Exploring dynamic self-adaptive populations in differential evolution

摘要: Although the Differential Evolution (DE) algorithm has been shown to be a simple yet powerful evolutionary for optimizing continuous functions, users are still faced with problem of preliminary testing and hand-tuning parameters prior commencing actual optimization process. As solution, self-adaptation found highly beneficial in automatically dynamically adjusting such as crossover rates mutation rates. In this paper, we present first attempt at self-adapting population size parameter addition Firstly, our main objective is demonstrate feasibility DE. Using De Jong's F1–F5 benchmark test problems, showed that DE self-adaptive populations produced competitive results compared conventional static populations. reducing number used DE, proposed actually outperformed one problems. It was also an absolute encoding methodology greater reliability relative methodology.