The improved Crown Porcupine Optimization Algorithm and its application

To enhance the search efficiency, convergence speed, and population diversity of the Crested Porcupine Optimization (CPO) algorithm, and to improve its global search capability. By introducing a direction-controlled golden sine strategy, the algorithm dynamically adjusts the search boundaries to quickly locate the global optimal solution. The parrot fear behavior is incorporated to optimize the local exploration and global search capabilities of the CPO. The optimization process was tested using 8 benchmark test functions, and the results, along with Wilcoxon signed-rank test results, demonstrated the superiority of the improved algorithm. Finally, the ICPO was applied to optimize key parameters of the VMD-LSTM prediction model for fault diagnosis in transmission lines, further validating the effectiveness of the improvement strategies and the superiority of the algorithm.