Multi-objective optimization and finite element method combined with optimization via Monte Carlo simulation in a stamping process under uncertainty

The response surface methodology (RSM), which uses a quadratic empirical function as an approximation to the original and allows identification of relationships between independent variables xi dependent ys associated with multiple responses, stands out. main contribution present study is propose innovative procedure for optimization experimental problems considers insertion uncertainties in coefficients obtained functions order adequately represent real situations. This new procedure, combines RSM finite element (FE) method Monte Carlo simulation (OvMCS), was applied stamping process Brazilian multinational automotive company. For were compared results using agglutination methods: compromise programming, desirability (DF), modified (MDF). optimized by applying generalized reduced gradient (GRG) algorithm, classic widely adopted this type problem, without uncertainty factors. advantages offered are presented discussed, well statistical validation its results. It can be highlighted, example, that proposed reduces, sometimes eliminates, need additional confirmation experiments, better adjustment factor values variable when comparing responses. added relevant useful information managers responsible studied process. Moreover, facilitates improvement process, lower costs.