Abstract:
In the optimization design of products and processes in the biological medicine, we need to consider multiple
characteristics of quality simultaneously, namely multi-response problems, multi-response optimization design can improve the quality of the products effectively, and realize enormous economic benefits and so multi-response optimization design is showing a more and more important role in continuous quality improvement activities. But usually there is no specific set of input variables to make all the response variables be optimal, and the traditional multiresponse surface method cannot solve the correlation problem between multi-responses and regression model problem effectively. Because we can make a better fitting model and solve the problem of the correlation between the response variables at the same time with SUR method, this thesis uses the SUR method to model the relationship between each response and control variables, and makes predictions; confirms the satisfaction function of each response and the overall satisfaction function; combines with practical problems of a company in biological medicine field named SX to conduct empirical research, this thesis confirms the optimal factor level combination with the overall satisfaction function in the end, thus solves the multi-response optimization problems.
