Including time dependency and ANOVA in decision-making using the revised fuzzy AHP: A case study on wafer fabrication process selection

In the process of decision-making, sometimes analysts are given a sub-set of criteria for which statistical data are available from experiments, whereas for some other criteria, only qualitative judgments can be made. In such situations, it is important to consider the weights of experimental criteria along with the decision maker's qualitative weights. Furthermore, in complex decision scenarios, one may need to consider the time notion and evaluate the behavior of alternatives over time, before making a final decision. This article using the revised fuzzy analytic hierarchy process (AHP) introduces a new decision process to include (1) time dependency of decisions and (2) statistical weighting from the standard analysis of variance (ANOVA). The application of the method is shown via a case study in the selection of wafer slicing and coating process for a three-year operation time. A signal-to-noise metric has been adapted to differentiate among alternatives that swap ranks over time. The method is straightforward and can be adapted to other multiple criteria decision making (MCDM) models.

► In this article we extend the revised fuzzy analytic hierarchy process to include time dependency.
► The method comprises both quantitative and qualitative criteria.
► For quantitative criteria, statistical weighting from ANOVA is incorporated.
► A signal-to-noise metric is adapted to make final decision for alternatives that swap ranks over time.