I have not heard about the same problem. There are two similar types of problems:
- A non-linear decrease in dimension, you are given N points with large sizes, and you want to find N low-dimensional points that preserve distance, as well as possible. MDS, mentioned by Michael Kowal, is one such method.
- : . , Kuhn-Munkres ( ), NxN, pi pj, . , , b- (Kuhn-Munkres 1-).
, " ", , (2), (2) , , .
, Kuhn-Munkres .