Optimizing Multi-Response Problems in the Taguchi Method by Fuzzy Multiple Attribute Decision Making

Lee-Ing Tong and Chao-Ton Su

Quality and Reliability Engineering International, Vol. 13, pp25-34, 1997

Traditionally, the Taugchi method is used in off-line quality control. One of the drawbacks to this technique is that is deals only with single-response problems. This paper deals with the Technique for Order Preference by Similarity to Ideal Solution (TOPSIS) which can handle multi-response problems with both continuous and discrete data. Theoretically, the TOPSIS indifference curve is very similiar to Taguchi's quadratic loss function.

"Multiple attribute decision making (MADM) invlolves the selection among some alternatives each having multiple, usually conflicting, attributes." TOPSIS looks for the alternative that is closest in distance to the ideal solution and furthest in distance from the negative-ideal solution. To calculate these distances, Tong and Su have six steps to follow. Once these six steps have been completed, one must then decide on the optimization procedure.

This consists of another six steps where fuzzy numbers are used. After walking the readers through this, Tong and Su present a case study to verify the optimization procedure they have described. This then leads them to the conclusions of the paper.

The following conclusions are reached by Tong and Su:

1) Each response's relative importance may be easily expressed with the linguistic term;

2) The TOPSIS value is the only one required for multiple responses at each experimental trial;

3) The procedure is universal in its approach and, therefore, may be used in any multi-response problem;

4) The TOPSIS method can simultaneously handle both discrete and continuous data.

For details on the above mentioned steps, please see the paper.

Arika Blankenship
December 2, 1997