http://www.daimi.au.dk/~doktoren/master_thesis/
-Josh
Neat site I came across for egtb.
-
- Posts: 489
- Joined: Sat Jan 21, 2006 10:43 am
- Sign-up code: 10159
- Location: Reading, UK
- Contact:
BDDs and Compression Results
This thesis investigates the use of Binary Decision Diagrams (BDDs) to see if they are a more compact and more efficiently-accessed way of storing chess endgame tables (EGTs).
The compression results seem, on aggregate, to be no better than achieved already with Nalimov EGTs.
The access-times do not really compare 'like with like' as the BDD access-times require the BDD to be (unscalably) in RAM, which a Nalimov EGT would not be.
BDDs are also at the centre of Joe Hurd's paper, producing the EGT results by theorem-proving techniques using HOL, q.v.
http://www.cl.cam.ac.uk/~jeh1004/resear ... /chess.pdf
g
The compression results seem, on aggregate, to be no better than achieved already with Nalimov EGTs.
The access-times do not really compare 'like with like' as the BDD access-times require the BDD to be (unscalably) in RAM, which a Nalimov EGT would not be.
BDDs are also at the centre of Joe Hurd's paper, producing the EGT results by theorem-proving techniques using HOL, q.v.
http://www.cl.cam.ac.uk/~jeh1004/resear ... /chess.pdf
g