Article: Integration of dynamic multi-response systems using the product of normalised squared-bias and variance Journal: International Journal of Quality Engineering and Technology (IJQET) 2012 Vol.3 No.2 pp.108 - 123 Abstract: The quality loss function developed by Taguchi for smaller-the-better and nominal-the-best characteristics can be decomposed into two components: variance and bias. For the larger-the-better characteristics with an infinite target, a decomposition of this nature is not straightforward. However, the decomposition for the larger-the-better characteristic is possible when a finite target is considered. This research proposes decomposing all three types of characteristics in the same way for similarity and equality among characteristics for multi-response cases. The two components of quality loss are utilised to formulate a metric called the product of normalised squared-bias and variance. This metric can in turn be employed to simultaneously improve both the dynamic and the static multiple responses of a system. An example is also provided to demonstrate the proposed methodology in a dynamic multi-response system. Inderscience Publishers - linking academia, business and industry through research

Title: Integration of dynamic multi-response systems using the product of normalised squared-bias and variance

Authors: Naresh K. Sharma; Elizabeth A. Cudney; Steven M. Corns

Addresses: Engineering Management and Systems Engineering, Missouri University of Science and Technology, 217 Engineering Management, 600 W. 14th St., Rolla, MO 65409-0370, USA. ' Engineering Management and Systems Engineering, Missouri University of Science and Technology, 217 Engineering Management, 600 W. 14th St., Rolla, MO 65409-0370, USA. ' Engineering Management and Systems Engineering, Missouri University of Science and Technology, 217 Engineering Management, 600 W. 14th St., Rolla, MO 65409-0370, USA

Abstract: The quality loss function developed by Taguchi for smaller-the-better and nominal-the-best characteristics can be decomposed into two components: variance and bias. For the larger-the-better characteristics with an infinite target, a decomposition of this nature is not straightforward. However, the decomposition for the larger-the-better characteristic is possible when a finite target is considered. This research proposes decomposing all three types of characteristics in the same way for similarity and equality among characteristics for multi-response cases. The two components of quality loss are utilised to formulate a metric called the product of normalised squared-bias and variance. This metric can in turn be employed to simultaneously improve both the dynamic and the static multiple responses of a system. An example is also provided to demonstrate the proposed methodology in a dynamic multi-response system.

Keywords: quality loss function; unified methodology; finite targets; signal-to-noise ratio; normalised squared-bias; variance; static systems; dynamic multi-response systems.

DOI: 10.1504/IJQET.2012.049682

International Journal of Quality Engineering and Technology, 2012 Vol.3 No.2, pp.108 - 123

Received: 23 Jan 2012
Accepted: 12 Mar 2012
Published online: 30 Aug 2014 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article

Title: Integration of dynamic multi-response systems using the product of normalised squared-bias and variance

Keep up-to-date