Which Alogorithms does Jess use?

[yadati chetan]

Does Jess use any algorithm for optimal processing of rules? if so are there any particular algorithms that it uses?
Specifically does jess use RETE algorithm?

[Frank Carver]

Jess definately does use the Rete algorithm. In fact one of the most important classes in the Jess rule-engine code is jess.Rete !

[Ernest Friedman-Hill]

Jess does indeed use the Rete algorithm for pattern-matching, with a number of optimizations and modifications. We've considered implementing alternate algorithms which could be switched on when desired: the TREAT and LEAPS algorithms would be the leading candidates. LEAPS is interesting because it makes the similarity between the working memory of a rule engine and the indexed store of a database more or less explicit; it also has the potential to be the fastest (there are no commercial implementations of either of these algorithms right now; they're both pretty tricky.)
[ September 04, 2003: Message edited by: Ernest Friedman-Hill ]

[Frank Carver]

the TREAT and LEAPS algorithms would be the leading candidates.
That's interesting. Do you have any references for further information on these algorithms?

[Ernest Friedman-Hill]

You can buy Daniel Miranker's Ph. D. thesis on Amazon for a discussion of TREAT. LEAPS was Don Batory's work; one reference is here but you really have to read a lot of papers to get enough information about it, because all of that group's writing is very self-referential and insular.
[ September 04, 2003: Message edited by: Ernest Friedman-Hill ]

[Frank Carver]

Just to note that the ftp URL you gave for the LEAPS paper seems unavailable. I was, however, able to find a PDF version of what I assume is the same paper at http://www.cs.utexas.edu/ftp/pub/predator/tr-94-28.pdf