Research output: Contribution to journal › Article
The paper is devoted to the problem of available bandwidth estimation for the network channels in distributed computing systems, in particular in learning distributed artificial neural nets. The linear model of estimation is described, four methods are proposed to solve the problem. The methods are compared by simulation, and the conclusion about the most appropriate method is derived.
| Original language | Russian |
|---|---|
| Pages (from-to) | 45-52 |
| Journal | НЕЙРОКОМПЬЮТЕРЫ: РАЗРАБОТКА, ПРИМЕНЕНИЕ |
| Issue number | 11 |
| State | Published - 2009 |
ID: 5014305