Selection and preprocessing of factors for simplex optimization

摘要: Abstract Factor selection and preprocessing are important aspects of optimization studies. When simplex is used, as many factors can be handled conveniently should included in the optimization; additional do not greatly increase number experiments required to reach optimum. If risk associated with poor response minimal, a large initial will provide information about larger volume factor space. It unnecessary assign each step size that inversely proportional factor's significance. The vertices easily calculated by minor modification published algorithm.