Improved Salp Swarm Algorithm for Tool Wear Prediction

To address the defects of salp swarm algorithm (SSA) such as slow convergence speed and ease falling into a local minimum, new combining chaotic mapping decay factor is proposed combined with back propagation (BP) neural network to achieve an effective prediction tool wear. Firstly, used enhance formation population, which facilitates iterative search reduces trapping in optimum; secondly, introduced improve update followers so that can be updated adaptively iterations, theoretical analysis validation improved SSA are carried out using benchmark test functions. Finally, strong optimization capability solve BP networks for optimal values hyperparameters used. The validity this verified by actual wear data set. results function show presented has better solution accuracy. Meanwhile, compared original algorithm, R2 value part life model from 0.962 0.989, MSE reduced 34.4 9.36, 72% improvement obtained.