3 ms·
Well, the simple approach is to use backtracking. From there you can dive into back jumping, constraint learning and DPLL. Now, implementing that is surely ful
by dietr1ch 3y ago
Well, the simple approach is to use backtracking. From there you can dive into back jumping, constraint learning and DPLL.
Now, implementing that is surely full of tricks too, which you may also find interesting, but is probably not discussed too much.
- inasio 3y agoThe simplest is probably WALKSAT-like heuristics, essentially local search: flip a random variable, evaluate, if you see an improvement accept the flip, otherwise try again. This can be surprisingly efficient on some SAT problems