An approach to rank MADM alternatives based on attribute state weights

To overcome the drawbacks that the existing alternative ranking approaches based on constant weights in multi-attribute decision making (MADM) rarely reflect the characteristics of complex systems such as nonlinear and emergence, an approach to rank MADM alternatives based on attribute state weights is presented. The presented approach relies on a new analysis structure constructed by the ideas incorporating reductionism with holism, in terms of which, attribute state weights are determined by solving a two-objective programming model that simultaneously minimizes subjective preference priorities and objective preference priorities on criteria cluster and on virtual goal cluster. Using the derived attribute state weights to compute the overall value for each alternative, the MADM alternatives are ranked. Besides the characteristics of complex systems are reflected, the presented approach also integrates subjective preference with objective preference, resulting in the presented approach is scientific and reasonable. By comparing with other approaches on a simulation example, the presented approach is proved to be scientific, reasonable, and well applicable to MADM alternative rankings.