Multi response optimization of Process parameters using weight Baseddesirability and grey relationAnalysis in the taguchi method

Abstract

Multiresponse optimization is an interesting area in which more newlinethan one factor is taken into account simultaneously to obtain an optimal newlinesolution for manufacturing problems Unfortunately it is difficult to optimize newlinea multiresponse problem by the Taguchi method In order to solve the multiresponse newlineproblem the multiple values are converted into single response value newlineand then optimal solutions obtained Until now engineering judgement has newlinebeen primarily used to solve such a complicated multiresponse problem But newlinethe method increases the uncertainty of the decision making process So newlineweight assigning must be scientific to avoid uncertainty newlineHow to determine a definite weight for each response in a real case newlinestill remains difficult Many researchers have proposed different methods of newlineassigning weights to responses for optimizing multiresponse problems newlineHowever their approach increases more complexity to the computational newlineprocess and the possible correlation among the responses may still not be newlineconsidered In this thesis Eigen valuebased weight is employed to find the newlineoptimal weights for each response so as to maximize the values of SN ratio newlineThe potential of these integrated approaches to solve the multiple newlineresponse problems has been demonstrated with four case studies Generally newlinein Grey relation analysis and desirability method the average values of grey newlinecoefficients are directly taken for calculating the grey grade This brings newlineuncertainty to the problem So there is a need for finding some scientific newlinemethod to get the optimal solution In the proposed method the grey newlinerelational grade is calculated based on the new weight which is multiplied newlinewith the existing grey relation coefficient for obtaining the grey grade which newlineis called WeightBased Grey Relational Analysis WBGRA newline newline