Advanced structural design of engineering components utilizing an artificial neural network and GNDO algorithm

Abstract In today’s competitive environment, the lightweighting of vehicle components is under intense study. While some of these studies focus on material modification, a very important part of these studies focuses on lightweighting the same material. The most widely used techniques in light-weight studies are topology, topography, size, shape optimization, and metaheuristic algorithms. This work introduces a novel hybrid generalized normal distribution optimization (GNDO) simulated annealing algorithm (GNDO-SA) adapted to optimize a vehicle component made of aluminum material. The focus is on shape optimization, which aims to minimize the weight of the vehicle component while ensuring that stress constraints are met. A combination of latin hypercube sampling (LHS) and artificial neural network is used to generate the mathematical equations governing mathematical equations for the objective/constraint used in the optimization. These findings highlight the effectiveness and superiority of the GNDO-SA method for optimization problems.