Robust Design of Product Based on Response Surface and Support Vector Machines

An approach to product robust design based on integrated surrogate models is proposed.Response surface and support vector machines are adopted to approximate the actual models in view of the quality characteristics of a product and the related design parameters,which ensures the accuracy and efficiency of constructing the surrogate models in the case of limited samples.To make a product’s quality insensitive to the noise and improve the product’s reliability,a percentile based optimization model for robust design is also presented by taking into account both the design objective robustness and the feasibility robustness of design constraints.A case study of a solar powered irrigation system is described to show the effectiveness and practicability of the proposed approach.