Abstract: In order to solve the problems of low precision, local optimization and slow convergence of subtraction average optimizer algorithm, a multi-strategy improvement method is proposed. Firstly, the population members are initialized using Logistic chaotic mapping to make the population distribution more uniform. Secondly, a nonlinear weight factor is introduced to reduce the influence of randomness on the optimization of the algorithm. Finally, the combination of variable spiral and Cauchy variation strategy updates the position of search agents, increases the diversity of particles, helps the algorithm to jump out of the local optimal region, explore a broader search space, improve computing efficiency, and find a balance between global search and local search. The performance of the improved VCSABO algorithm is tested and evaluated by eight benchmark test functions. The experimental results show that the convergence speed, calculation accuracy and optimization ability of the improved subtraction average optimizer algorithm are greatly improved, and the effectiveness of the improved strategy is verified. At the same time, VCSABO algorithm is used to optimize the variational mode decomposition algorithm to process CWRU fault signals. The results show that the key parameter combination (K, α) can be obtained adaptively by this method, and the characteristic components in the signal can be effectively decomposed, which shows significant technical advantages in the processing of weak fault signals of rotating machinery and validates the engineering application value of VCSABO algorithm.
