An empirical study on Differential Evolution algorithm and its several variants

Differential Evolution (DE) is a simple and efficient global optimization algorithm. However, DE has indicated its weaknesses, such as the convergence rate. This fact has inspired many computer scientists to improve upon DE by proposing modifications to the original algorithm. This paper presents a survey on DE and its several variants. In this paper, we design two sets of function optimization experiments. One is about the five different mutation strategy of the Conventional Differential Evolution (CDE), and the other is the comparison several variants of Differential Evolution algorithm with a new improved DE algorithm (GPBXDE). To evaluate the performance of the algorithm, we selected twelve widely used benchmark functions. The results of the experiment prove that the strategy CDE/rand/1/bin and CDE/rand-to-best/1/bin are better and the GPBXDE algorithm performs outstanding.