Premium
OPTIMIZATION THROUGH EXPERIMENTATION: APPLYING RESPONSE SURFACE METHODOLOGY
Author(s) -
Brightman Harvey J.
Publication year - 1978
Publication title -
decision sciences
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.238
H-Index - 108
eISSN - 1540-5915
pISSN - 0011-7315
DOI - 10.1111/j.1540-5915.1978.tb00737.x
Subject(s) - response surface methodology , computer science , context (archaeology) , mathematical optimization , process (computing) , design of experiments , inference , mathematics , machine learning , artificial intelligence , paleontology , biology , operating system , statistics
The crucial steps in a quantitative analysis of a decision problem are problem formulation, model building, analysis, and implementation. Given an initial model specification, the goal of analysis is to determine the values of the controllable or decision variables that optimize the objective function. Frequently the initial model is inadequate and must be reformulated. While modeling is an evolutionary process involving art and science, under certain conditions Response Surface Methodology (RSM) is an effective vehicle for constructing and parameterizing optimization models. RSM, which draws upon the areas of experimental design, modeling, inference, and optimization, utilizes different opening and ending strategies. Through simultaneous and sequential experimentation, the approximate region of the model's maximum response is found by employing the steepest ascent method. Subsequently, the exact values of the controllable variables that maximize the model's response are determined by canonical analysis. The RSM concepts are first developed within the context of a manufacturing problem. A potential application to simulation studies is then presented.