摘要
This paper introduces two novel concepts, the one-variable update mechanism (UM1) and the harmonic step-length strategy (HSS), to improve the continuous simplified swarm optimization (SSO). The proposed UM1 and HSS are able to balance the exploration and exploitation ability of the continuous SSO in solving high-dimensional multivariable and multimodal numerical continuous benchmark functions. The proposed new update mechanism UM1 updates only one variable and it is total completely different to that in the SSO without needing to update all variables. In the proposed UM1, the HSS enhances exploitation capacity by decreasing the step-length based on a harmonic sequence. Numerical experiments are performed on 18 high-dimension functions to confirm the efficiency of the proposed approach.
•A new continuous simplified swarm optimization (SSO) is proposed.•The proposed new SSO implements the one-variable update mechanism (UM1) and the harmonic step-length strategy (HSS).•The proposed UM1 updates one variable in each solution to improve the exploration ability.•The proposed HSS enhances exploitation ability by reducing the length of the search area from long to short based on a harmonic sequence.