A Multidimensional Scaling (MDS) method for interval-valued dissimilarity data. The objects are represented as hyperspheres in a p-dimensional space. The minimum and maximum distances between hyperspheres approximate the lower and upper dissimilarities in the least squares sense. The method is described in:
T. Denoeux and M.-H. Masson. Multidimensional scaling of interval-valued dissimilarity data. Pattern Recognition Letters, 21:83-92, 2000. pdf