Using the GA Package in R Program and Desirability Function to Develop a Multiple Response Optimization Procedure in Case of Two Responses

  • Saranya Thongsook
Keywords: Response surface methodology, statistical software packages, desirability function


Multiple responses are one of the most frequently used methods in present experimental works. The desirability function is one popular approach for multiple response optimization (MRO). Many statistical software packages can also be used as optimization software for desirability function such as Design Expert, MINITAB, SAS, R and so forth. Most distributed software has more restrictive aspects governed by software licenses, while R is a free software that includes several optimization methods. The purpose of this research is to develop MRO procedure via the GA package in R and the desirability function to find a global optimization of multiple responses in case of two responses. The results from the GA package were compared with the results from the optim function, the rgenoud package, the DEoptim package and Design Expert (v.9 (Trial Version)) based on other algorithms to illustrate its performance. The results showed that the best values of three GA package methods were equal or superior to those values of other methods in R. Moreover, they were superior to those values of Design Expert in some cases. Therefore, it concluded that the GA package methods using built-in standard genetic operators in R and desirability function is a suitable method for finding a global optimization of multiple responses.