BNMSBritish Nuclear Medicine Society
BNMSBelize National Meteorological Service
BNMSBroadband Network Management System
BNMSBearer Node Management System
Copyright 1988-2018, All rights reserved.
References in periodicals archive ?
[3] treated the routing changing as noise interference, by adding transition probabilities in static BNM, which improved the inference precision.
In BNM of IP network, the evidence nodes are the state variables of E2E path properties, the hidden nodes are the state variables of link properties, and the directed edges represent the relationships among the E2E paths and corresponding traversing links.
We can remove the evidence node [y.sub.i] and its corresponding hidden nodes [x.sub.j], as well as the directed edges connecting them of BNM. So, we make Definition 2 as follows.
BNM of IP network will be converted into Rest BNM (RBNM) after removing evidence nodes of [y.sub.i] = 0 (good E2E paths), hidden nodes (links traversed by good E2E paths), and corresponding directional edges.
When inferring the set of congested links under BNM, according to the dynamic routing algorithm strategy in IP network, for N times of E2E path snapshots, links traversed by E2E paths will be changed due to some factors, such as the physical link cut or link bandwidth limited.
If the structure (including parameter) of a discrete static BNM will change in different time, this kind of model is called Variable Structure Discrete Dynamic Bayesian (VSDDB) network model [19].
E2E paths and corresponding traversing links in the present inference moment can be obtained from the congested BNM in the time slice [T.sup.2] after removing nodes and directed edges representing good paths and corresponding traversing links; the RBNM and corresponding rest congested matrix of IP network in the present inference moment can be obtained.
The state variables of congested path are [mathematical expression not reproducible] (evidence nodes), and the state variables of links [mathematical expression not reproducible] (hidden nodes) in each piece of time slice build the corresponding static BNM. Each static BNM in the two time slices constitutes VSDDB network model through the node interfaces.
From the stored BNMs at each node, individual unit-controllers compute the best routes to reach units contained in the destinations.
This information is encapsulated within the BNM. The information encapsulated within the BNM provides information regarding the inter-connections between the brokers in the cluster that it is a part of, the interconnections between the clusters within the super-cluster that it belongs to and so on.