An interactive sorting approach based on information theoretic measure

In this study, we develop an interactive approach for sorting alternatives. We assume that the preferences of the decision maker are consistent with an additive function. We assign worst and best possible categories for each alternative and narrow down these category ranges using mixed integer programming (MIP) iteratively. We utilize binary variables to assign the alternatives for which the classes are not known exactly. We incorporate the worst and best possible category information to the MIPs whenever new information is obtained. In each iteration, we find the assignment frequency of alternatives for each category using the values of the binary variables gathered from the MIPs and calculate the probability of an alternative to be placed in a category based on these frequencies. We then use the information theoretic measure, entropy, in the selection of the alternative that will be placed to a category by the decision maker. The entropy concept fits well to our measurement of uncertainty about the categories of the alternatives since the aim is to ask the decision maker an alternative that will yield much information to the decision process. We test the performance of our approach on an example problem from the literature and the results show that it works well.
Citation Formats
A. Özarslan and G. Karakaya, “An interactive sorting approach based on information theoretic measure,” presented at the 25th International Conference on Multiple Criteria Decision Making (16 - 21 Haziran 2019), İstanbul, Türkiye, 2019, Accessed: 00, 2021. [Online]. Available: