SAT Flashcards

1
Q

describe chronological backtracking?

> performance?

A

chronological backtracking: backtracking to the highest decision level that has not been tried with both values
> performs okay for randomly generated instances, performs bad for instances in practical applications

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
2
Q

describe conflict driven learning

> what is the advantage/disadvantage?

A

conflict driven learning: solve until you reach a conflict
> formulate conflict as clause and add to list of clauses
> advantage: solver will not get into the same conflict again

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
3
Q

what is decidability?

A

decidability: the guarantee to find an answer/find that there is no answer (given enough memory/runtime)

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
4
Q

what are sources of incompleteness?

A

> incompleteness of input data

> insufficient runtime to wait for answer

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
5
Q

what is undecidability?

A

undecidability: not always guaranteed to find an answer

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
6
Q

which SAT problems are hard?

A

with n = number of variables and l = number of clauses

l/n = 4.26 > hard SAT problem

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
7
Q

what are the properties of the SAT-complexity peak?

A
  1. very stable across problem size and solver types
  2. coincides with solubility transition
    > l/n < 4.3 problems are underconstrained and SAT
    > l/n > 4.3 problems are overconstrained and UNSAT
    > this is also called “phase transition”
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
8
Q

how does a local SAT search work?

A

local search:
1. make a random guess about variable values
2. test how wellthe guess is doing
3. flip a variable to improve
> e.g. flip variable that increases the number of clauses satisfied

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
9
Q

local search: does flipping the most improving variable work?

A

> YES for 2SAT

> NO in general: usually does get stuck locally

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
10
Q

explain GSAT

A

greedy SAT: because local search gets stuck, restart periodically with new initial guess

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
11
Q

what is unusual about the search behavior of GSAT?

A

a lot of time is spent searching sideways > mostly searching on plateaus

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
12
Q

what is MAXSAT?

A

MAXSAT is an optimization extension of SAT

> goal: find the maximum number of clauses that can be satisfied simultaneously

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
13
Q

MAXSAT: how to model clause relevance?

A

model by assigning a cost to each clause if UNSAT
> if clause C is mandatory, cost is infinite and C is called HARD
> otherwise C is called SOFT

> any truth assignment has a cost which is the sum of all clause costs. try finding the truth assignment with the lowest finite cost

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
14
Q

name 4 MAXSAT categories

A
  1. MAXSAT (ms): no hard clauses and all clauses have weight 1
  2. weighted MAXSAT (wms): no hard clauses
  3. partial MAXSAT(pms): hard clauses but all soft clauses have weight 1
  4. weighted partial MAXSAT (pwms): hard clauses and individual weight of soft clauses
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
15
Q

what is 2SAT? (or 3SAT)

A

2SAT: conjunction of disjunctions of length 2

> solvable in polynomial time

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
16
Q

what are horn clauses?

A

horn clause: a clause with max 1 positive literal