Multi-objective optimization of robust parameter design based on Gaussian process model

In engineering design, robust parameter design (RPD) is commonly employed to optimize problems that involve noise factors. Bayesian optimization algorithms have gained increasing attention in RPD, particularly in practical production settings where multi-objective optimization problems are frequently encountered. This article utilizes a Gaussian process model to simulate the response function and integrates it with robust parameter design. Through multi-objective optimization algorithms, the optimal solution set is obtained, leading to the determination of the optimal design scheme. Both control factors and noise factors in quality design are taken into consideration. Building upon the Bayesian optimization algorithm, a new optimization criterion is proposed, which effectively reduces the number of iterations, minimizes computational costs, and enhances computational accuracy.