| S. Hanly and D. Tse, "Not known yet." Check with Stephen and David, 1995. |
....; Y jX S ) I(X S ; Y ) The last step in the above inequality follows from the chain rule I(XL ; Y ) I(XM ; Y ) I(X LnM ; Y jXM ) Chain Rule) 3 A few basic facts on convex polytopes may be found in Section 3B. 4 This follows from the fact that, as observed by Hanly, Tse, and Whiting [16, 17], R is actually a convex polytope of a special kind called bounded polymatroid. One of the properties of bounded polymatroids is that they have a face that dominates every point. which holds for any M L [M ] Notice that if X 1 ; Delta Delta Delta ; XM are independent random variables ....
.... i 2 [M ] gets assigned, where (i) 1 ( i) M) corresponds to the position at the very top (bottom) In this way, each 2 Pi defines a rate tuple R 2 D with R i = I(X i ; Y j [ j2 Gamma1 (f1; Delta Delta Delta; i) Gamma1g) X j ) 38) It was observed by Hanly, Tse, and Whiting [16, 17] that R is in fact a bounded polymatroid. Polymatroids were introduced by Edmonds [20] who also derived their basic properties. For an introduction into polymatroid theory see, e.g. 21] The following lemma is a direct consequence of the fact that R is a bounded polymatroid. Lemma 14 (See ....
S. Hanly and D. Tse, "Not known yet." Check with Stephen and David, 1995.
Online articles have much greater impact More about CiteSeer.IST Add search form to your site Submit documents Feedback
CiteSeer.IST - Copyright Penn State and NEC