Talk:Backjumping

NCB/DDB
Is this the same thing as nonchronological or dependency-directed backtracking?
 * They are related, though they have slightly different nuances. As far as I can recall, the main difference is in how they record the relations between the current and previous variables in order to backtrack. Diego (talk) 05:46, 19 August 2008 (UTC)

Applications
In what applications is this technique actually used? -- Beland (talk) 17:51, 17 August 2008 (UTC)
 * Almost any modern automatic problem solver based on branch and bound must use a variant of this technique. Diego (talk) 05:49, 19 August 2008 (UTC)

Clarity
This really needs cleaning up. Some of the explanations here are very complicated - at least, moreso than they need to be. I'd also question the bit which says that conflict-directed backtracking was invented in 1993. I'm almost certain this stuff has been around since the 1970s. —Preceding unsigned comment added by 86.189.4.43 (talk) 16:28, 3 January 2010 (UTC)
 * Do you have some references, or at least some idea of which field had something similar? Prosser 93 is well known as a seminal work in the context of CSPs. I don't doubt similar ideas were tried earlier, but I don't know any earlier paper where problems were described in the abstract CSP form and direct conflicts were recorded during search. Also, it might be that the technique you know was slightly different than Prosser's, which is the one giving its name to CBJ. Diego (talk) 19:49, 3 January 2010 (UTC)

Gashnik's phd thesis was the first attempt at conflict-directed backjumping (but only from leaf nodes) AFAIK 213.174.214.36 (talk) 17:13, 8 September 2017 (UTC)