| 摘要 |
In a local area network emulation architecture, an improved search approach is achieved when the set of nodes to be searched is divided into N subsets for a seriatim search through the subsets. The allocation of the nodes to the different subsets is algorithmically determined for minimum, or near minimum, utilization of resources. Illustratively, the algorithm disclosed determines whether a node in subset i should be reassigned to subset i+1 based on whether where epsilon is the probability that the node under consideration contains the information searched for, pi is the probability that the information searched for is found in subset i, and ki+1 is the number of nodes in subset i+1. The algorithm determines whether a node in subset i should be reassigned to subset i-1 based on whether where epsilon is the probability that the node under consideration contains the information searched for, pi-1 is the probability that the information searched for is found in subset i-1, and ki is the number of nodes in subset i. As time progresses and search results are accumulated, the results are used to reassess probability values and to reapply the reassignment thresholds.
|
