A new method for weighting decision making attributes: an application in high-tech selection in oil and gas industry

In multiple attribute decision making (MADM), there is often a need for comparing the attributes. In situations such as time pressure and lack of adequate knowledge, eliciting exact numerical weights from decision maker (DM) is impossible, but instead, obtaining a ranking of the attributes using an elicitation method is practical. Hence, many investigators have concentrated on a branch of the methods called approximate weighting. These methods generate the surrogate weights for the precise numbers that cannot be extracted from the DM’s mind. Many researchers have attempted to get the best approximate weights as close as possible to precise weights. There are several approximate weighting methods; among others, it is well known that the rank order centroid (ROC) outperforms the others. In competition with the ROC method, this paper develops a novel method, called rank order logarithm (ROL), that is based on a justifiable and well-founded concept. Three evaluations are performed to compare the ROC and ROL methods: a set of simulation experiments, a theoretical analysis, and a real-case analysis. Interestingly, the results of all the three evaluations reveal that the ROL method is comparable to the ROC method in most situations. To show applicability of the ROL method in real-world situations, a study case taken from high-technology selection in petroleum industry is examined.