Robust parameter design of latent variable Gaussian process model based on mixed input variables

As for the problem of robust parameter design with qualitative and quantitative mixed input variables in the additive manufacturing process, a new optimization scheme is proposed based on the latent variable Gaussian process model. Firstly, the experimental data of mixed input variables are generated by maximum projection design. The part entities are obtained by 3D printing experiments. The response values of the target quality characteristics are collected by experiments. Then, each qualitative factor is mapped to potential numerical variables, and a Gaussian process model of mixed input latent variables i s constructed. Finally, considering the uncertainty of response, an optimization model is established, and the differential evolution algorithm is used to find the best factor design in the parameter space. The paper show that the method can scientifically express the relationship between qualitative factors. At the same time, it can effectively deal with the problems of meta-modeling and robust parameter design with mixed input variables.