Special issue on the dempster-shafer theory of evidence: an introduction
Liping Liu · International Journal of Intelligent Systems · 2003
The notion of belief functions is a result of the seminar work of Shafer 1 and its conceptual forerunner-lower and upper probabilities.2,3 In recent years, we saw a surge of interest in advancing this theory, developing efficient computation algorithms, and applying it to a wide range of business and engineering problems.Among many theoretical advances, the most noticeable are the theory of linear belief functions, 4 the theory of transferable beliefs, 5 and the theory of hints.6 The Dempster-Shafer theory of belief functions has become a primary tool for knowledge representation that bridges fuzzy logic and probabilistic reasoning.It has been applied to many areas including business decision making, accounting, finance, production, management information systems, and electronic commerce (e.g., see Ref. 7).To foster the exchange of these advances and applications, the Institute for Operations Research and Management Sciences invited over 30 scholars to present their current research in a cluster of seven regular sessions and one tutorial at its 2001 annual meeting held at Miami Beach, Florida.The topics of the presentations included: Belief Functions in Expert Systems; Belief Functions in Business Decisions; Belief Functions in Management Information Systems; Belief Functions and Theory of Hints; Belief Functions and Statistics; Modeling Using Belief Functions; and Belief Functions in Electronic Commerce.Arthur P. Dempster and Paul-Andre ´ Monney presented the tutorial to the general audience of the meeting.As per the requests of many participants and in collaboration with independent reviewers and Ronald Yager, the editor of the International Journal of Intelligent Systems, we have selected 7 articles from the 21 presentations to publish in this special issue.To reflect the theme of the meeting, the selected articles fairly cover three dynamic areas of ongoing research on belief functions, namely applications, theoretical advances, and computational implementation.The following paragraphs summarize the selected papers in the order of their presentations in this issue.