INVESTIGADORES
MOLTER ursula Maria
artículos
Título:
Computing the Hutchinson distance by network flow methods
Autor/es:
BARAHONA, FRANCISCO; CABRELLI, CARLOS A.; MOLTER, URSULA M.
Revista:
RANDOM & COMPUTATIONAL DYNAMICS
Editorial:
Dekker
Referencias:
Lugar: New York; Año: 1992 vol. 1 p. 117 - 129
ISSN:
1061-835X
Resumen:
We show, that the Hutchinson distance between discrete probability measures on spaces of finite dimension can be computed as a network flow problem, and then solved in polynomial time by efficient combinatorial mehtods. In the one dimensional case, a linear time method had been obtained earlier by Brandt et. al in [12]. This new setting allows to compute this metric for higher dimensions. We also show that the one dimensional formula hs its counterpart in the continuous case.
