Generalization of stable preference ordering towards ideal solution approach for working with imprecise data

When solving real-world decision-making problems, it is important to deal with imprecise quantitative values modeled by numerical intervals. Although a different extension of the multi-criteria decision-making methods could deal with intervals, many of them are complex and lack such properties as robustness to rank reversal. We present an extension of the stable preference ordering towards ideal solution (SPOTIS) rank reversal free method to deal with imprecise data. This extension of SPOTIS is also rank reversal-free. It offers a new efficient approach for solving multi-criteria decision-analysis problems under imprecision and can use different metrics of distance between intervals. The proposed approach is compared to the popular Interval technique for order preference by similarity to ideal solution) extension and performs very similarly to it. We also show on a practical example that the interval TOPSIS approach is not robust to rank reversal, contrary to our new SPOTIS extension approach, which offers a stable decision-making behaviour.