Quality evaluation model using loss function for multiple S-type quality characteristics

Abstract

Most of the studies of quality system or productquality assessment deal with a single quality characteristic to determine the quality loss. Products are often assessed on more than one quality characteristic. For this reason, different multivariate quality loss functions have been proposed. However, these loss functions only consider the nominal-the-best quality characteristics (N-type); they do not consider the condition when the quality characteristics are of the smaller-the-better (S-type).In this article, we present a quality evaluation model using loss function for multiple S-type quality characteristics. A numerical example is illustrated showing that using inappropriate loss functions will lead to inaccurate results that give either an underestimate or overestimate of the expected quality costs.

Achieving cost robustness in processes with mixed multiple quality characteristics and dynamic variability

“…Pignatiello (1993) considered a multivariate loss function to measure quality based on the interactions among multiple N-type characteristics; Tsui (1999) extended this work to include S-and L-type characteristics. Many others have continued to explore this area, including Kapur and Cho (1996), Teeravaraprug and Cho (2002), and Chan et al ( , 2005. The amassed effects of both univariate and multivariate quality loss efforts ultimately have resulted in the integration of the customer's perspective into the process mean problem, marking a significant shift from the traditional focus on meeting the manufacturers' needs in conforming to specifications.…”

Optimization of engineering tolerance design using revised loss functions

“…However, Cho and Phillips (1998) stated that the gamma distribution is more appropriate for the 'smaller-the-better' (S-type) quality characteristic and Chan et al (2005) extended the cost model using loss functions for multiplying S-type quality characteristics. Research on the economic tolerance design under a non-normal distribution on the 'nominal the best' (N-type) quality characteristics needs to be further explored.…”

Research on Quality Supervision Model for Cooperative Manufacturing Preparing Process in Virtual Enterprise

“…Thirdly, the asymmetry of quality information in manufacturing process of suppliers is fully considered and the issues of quality appraisal in outsourcing and transfer payment under the asymmetric information are mainly studied by establishing the principalagent model of manufacturing outsourcing; and then, the optimal solution of transfer payment is proposed by maximum principle [6]. Fourthly, a comprehensive hierarchical appraisal model of complex environment system based on the method of index weighting of subjective and objective is constructed by using empirical research method [7][8], and the concept of multivariate quality loss is proposed; and then, the model of multiple S-type quality loss appraisal model is designed [9]. Fifthly, the model of intelligent quality appraisal and quality prediction which are suitable for small-batch production process is presented, which gives corresponding prediction process and algorithm, achieves the goal of prediction effect optimizing the selection of samples by blurring samples with membership function, and avoids the disadvantages of overlearning and lower-generalization ability of intelligent methods such as artificial neural network, etc.…”