Multi-Objective Optimization Design of Vehicle Side Crashworthiness Based on Machine Learning Point-Adding Method

Multi-objective optimization problems are often accompanied by complex black-box functions which not only increases the difficulty of solving, but also solving time. In order to reduce computational cost such multi-objective problems, this paper proposes an ARBF-MLPA (Adaptive Radial Basis Function neural network combined with Machine Learning Point Adding) method, uses ABRF Function) and OLHS (Optimized Latin Hypercube Sampling) establish first generation metamodel NSGA-II (Non-dominated Sorting Genetic Algorithm II) algorithm obtain optimal front edge Pareto. The method is continuously used select add points update meta-model, then dynamically improve accuracy meta-model until converges. Then RBF-UDPA (Radial Uniform compared using test three different frontier features. performance evaluation indexes Inverted Generation Distance (IGD), Hypervolume (HV) Spacing Metric superior RBF-UDPA. Finally, applied in design vehicle-side crashworthiness. model converges after 6 iterations. Comparing results obtained finite element simulation results, error within 5%, meets requirements. optimized reduces chest intrusion 4.32%, peak collision force 2.11% mass 14.05%.