A dimension-selection based constrained multi-objective optimization algorithm using a combination of AI methods

Abstract The computational cost of modern simulation-based optimization tends to be prohibitive in practice. Complex design problems often involve expensive constraints evaluated through Finite Element Analysis or other computationally intensive procedures. To speed up the process and deal with constraints, a new dimension-selection based constrained multi-objective (MOO) algorithm is developed combining LASSO regression, artificial neural networks, Grey Wolf Optimizer, named L-ANN-GWO. Instead considering all variables at each iteration during optimization, proposed only adaptively retains that are highly influential on objectives. un-selected adjusted satisfy local search. With numerical benchmark engineering problem, L-ANN-GWO outperforms state-of-the-art MOO algorithms. method then applied solve complex high-temperature superconducting magnet. optimal solution shows significant improvement as compared baseline design.