1. Canonical Decomposition of Basic Belief Assignment for Decision-Making Support
- Author
-
Florentin Smarandache, Jean Dezert, DTIS, ONERA, Université Paris Saclay [Palaiseau], ONERA-Université Paris-Saclay, UNIVERSITY OF NEW MEXICO GALLUP USA, Partenaires IRSTEA, and Institut national de recherche en sciences et technologies pour l'environnement et l'agriculture (IRSTEA)-Institut national de recherche en sciences et technologies pour l'environnement et l'agriculture (IRSTEA)
- Subjects
Exponential complexity ,[PHYS]Physics [physics] ,Canonical decomposition ,Theoretical computer science ,Computer science ,Basic belief ,media_common.quotation_subject ,PCR5 ,020206 networking & telecommunications ,02 engineering and technology ,16. Peace & justice ,[SPI]Engineering Sciences [physics] ,Belief functions ,0202 electrical engineering, electronic engineering, information engineering ,Fuse (electrical) ,Decomposition (computer science) ,Decomposition method (queueing theory) ,020201 artificial intelligence & image processing ,Quality (business) ,Support system ,[INFO]Computer Science [cs] ,[MATH]Mathematics [math] ,media_common ,Decision-making - Abstract
International audience; We present a new methodology for decision-making support based on belief functions thanks to a new theoretical canonical decomposition of dichotomous basic belief assignments (BBAs) that has been developed recently. This decomposition based on proportional conflict redistribution rule no 5 (PCR5) always exists and is unique. This new PCR5-based decomposition method circumvents the exponential complexity of the direct fusion of BBAs with PCR5 rule and it allows to fuse quickly many sources of evidences. The method we propose in this paper provides both a decision and an estimation of the quality of the decision made, which is appealing for decision-making support systems.
- Published
- 2020