An enhanced crow search algorithm based on the disturbance method for engineering design problems

To address the shortcomings of the basic crow search algorithm (CSA), an enhanced crow search algorithm based on the disturbance method (DMCSA) is proposed. The basic CSA has slightly weak exploration and development abilities. The fixed flight length causes the basic CSA to conduct intensive searches in small areas, and it cannot conduct comprehensive and broad searches in the entire search space. The trigonometric function updating strategy is adopted to obtain a random walk model, so the global search ability is strong in DMCSA. In addition, the disturbance searching operation is adopted for detailed local search in DMCSA, as it can simultaneously balance convergence and diversity in the search space. Finally, this paper tests basic functions and multimodal functions and also provides iteration numbers, box plots, search paths, radar charts, the Wilcoxon rank sum test, graphs, and confidence intervals to comprehensively show different aspects of the performance of DMCSA in different functions. The experimental results show that the DMCSA efficiency reached 100% for all the high-dimensional functions. Among the two-dimensional functions, the worst-case function efficiency was 40%, and more than 60% efficiency was reached for all the other functions. For DMCSA application, this paper uses three engineering design problems to show different efficiencies. All results show that the DMCSA has excellent balance, good accuracy, wide exploration, and a high iteration speed.