|
RESEARCH INTERESTS |
|
For a manufacturing process with multiple response variables, it is desirable to find a best combination of input process variables that ensures every response variable is close to its target value and with minimum variability around the target value. We have proposed a new method that integrates multiple regression technique and Taguchi's signal-to-noise (SN) ratio concept. It is observed that the proposed method is superior to other methods with respect to total SN ratio as well as closeness of individual responses to their respective target values. [Quality Engineering, 22(4), 336-350, 2010]. We have also compared the effectiveness of weighted signal-to-noise ratio (WSN) as performance measure with other performance measures through the analysis of few published experimental data. [Computers & Industrial Engineering, 59(4). 976-985, 2010]. We have proposed similar multiple regression based modelling approach for optimization of dynamic multiple response variables. In a dynamic system, the target value of any response variable depends on the input signal set by the system operator. That means, there are multiple target values of the response variable depending on the setting of signal variables of the system. The proposed approach is shown as superior to other methods with respect to the closeness of individual responses to their respective target values and the expected variability associated with them. We are now working on methods for optimization of multiple responses where one or more response variables are categorical in nature. Copula-based regression models for a bi-variate mixed continuous and discrete variables are tried. Few utility functions that can be used for simultaneous optimization of both types of response variables are considered. These functions are being examined thoroughly using various datasets. |
|