COMPARATIVE ANALYSIS OF THE FUNCTION AND PACKAGES IN R FOR MULTIPLE RESPONSES IN CASE OF TWO RESPONSES

Abstract

Global Optimization is an important past in response surface methodology (RSM). Statistical package program is mostly used to find the global optimization such as statistical analysis system (SAS), statistical package for the social science (SPSS), MINITAB, DESIGN EXPERT, JMP, R and so forth. However, most softwares were restricted as they were a licensed software for example Design Expert, Minitab, SAS, SPSS and so forth. On the other hand, R is a free software that also includes many functions and packages to find the global optimization. However, its results should be carefully applied. Moreover, the application of R for the global optimization in case of multiple responses in RSM has never been investigated. Therefore, the purpose of this research is to analytical compare the function “optim”, the package “rgenoud” and the package “DEoptim” in R program version 3.2.3 for the global optimization in case of multiple responses in RSM by using the maximum result of overall desirability obtained from the function and the packages.  The result of this research concludes that the function “optim” by using Nelder and Mead method is an appropriate method to find the global optimization in case of multiple responses in RSM by consideration the maximum result of overall desirability.