A parallel approach to evidence combination on qualitative Markov trees

Author

Hong, Xin ; Liu, Weiru ; Adamson, Kenny

Author_Institution

Sch. of Comput. & Inf. Eng., Ulster Univ., Coleraine, UK

fYear

2003

fDate

27-29 Aug. 2003

Firstpage

522

Lastpage

526

Abstract

Dempster´s rule of evidence combination is computational expensive. We present a parallel approach to evidence combination on a qualitative Markov tree. Binarization algorithm transforms a qualitative Markov tree into a binary tree based on the computational workload in nodes for an exact implementation of evidence combination. A binary tree is then partitioned into clusters with each cluster being assigned to a processor in a parallel environment. The parallel implementation improves the computational efficiency of evidence combination.

Keywords

Markov processes; computational complexity; inference mechanisms; parallel algorithms; trees (mathematics); uncertainty handling; Dempster rule; binarization algorithm; binary tree; computational efficiency; computational work load; evidence combination; parallel processing; partitioned clusters; qualitative Markov trees; Bayesian methods; Binary trees; Clustering algorithms; Computational complexity; Computational efficiency; Computer architecture; Computer networks; Distributed computing; Parallel processing; Partitioning algorithms;

fLanguage

English

Publisher

ieee

Conference_Titel

Parallel and Distributed Computing, Applications and Technologies, 2003. PDCAT'2003. Proceedings of the Fourth International Conference on

Print_ISBN

0-7803-7840-7

Type

conf

DOI

10.1109/PDCAT.2003.1236357

Filename

1236357