A preference-preserving projection technique for MCDM

The paper introduces a new method for the graphical representation of multicriteria decision problems with discrete sets of alternatives. The approach proposed here extends projection techniques based on principal component analysis by incorporating information about preferences between alternatives in the graphical display. The advantages of principal component projection, which provides a clear view of data values, and a projection orthogonal to indifference hyperplanes, which provides preference information, are both made available to the user through a linear combination of projection matrices with interactively changed weights. The approach developed here can further be extended to more complex, nonlinear preference structures, and strategies for that extension are also discussed in the paper.