Improved Multi-objective Particle Swarm Algorithm Combined with North Goshawk Optimization Hyperparametric Optimization Least Squares Support Vector Machines for Linear Motors

In this study, we propose a Least Squares Support Vector Machine (LSSVM) combining the hyperparametric optimization of the Northern Goshawk algorithm (NGO) with an improved Multi-Objective Particle Swarm Optimization (MOPSO) Algorithm for the optimization of linear motors. Meanwhile, in order to improve the performance of the MOPSO algorithm in terms of population diversity, convergence speed and local search ability, we incorporate chaotic mapping and dynamic weighting mechanism. With this approach, the performance of linear motors can be predicted more accurately and their design parameters can be optimized. Firstly, a finite element simulation model of the motor is constructed, and analytical models for thrust force and thrust fluctuation are theoretically derived. Secondly, the Plackett-Burman design is employed to screen key factors and establish an experimental design space, reducing unnecessary variables and improving optimization efficiency. Subsequently, predictive regression modeling of the experimental space is performed using LSSVM, with hyperparameters optimized by NGO to enhance model performance. Finally, an iterative optimization of the combined model is conducted using the improved MOPSO to identify optimal structural parameter configurations. The optimized results are verified through finite element analysis (FEA), confirming the effectiveness and accuracy of the proposed method in enhancing motor performance