Constraint Satisfaction Problems

In which we see how treating states as more than just little black
boxes leads to the invention of a range of powerful new search
methods and a deeper understanding of problem structure and
complexity.

1
 Outline
• Constraint Satisfaction Problems (CSP)
• Backtracking search for CSPs
• AC-3 algorithm
• Local search for CSPs

2
 What is Search For?
• Planning: sequences of actions
– The path to the goal is the important thing
– Paths have various costs, depths
– Heuristics give problem-specific guidance
• Identification: assignments to variables
– The goal itself is important, not the path
– All paths at the same depth (for some formulations)
• CSP is a special class of search problems
– Mostly identification problems
– Have specialized algorithms for them
3
 State Representation

B C

B C
(a) Atomic (b) Factored (c) Structured

a) Atomic representation: a state is a black box with no internal structure;

b) Factored representation: a state consists of a vector of attribute values;
values can be Boolean, real-valued, or one of a fixed set of symbols.
c) Structured representation: a state includes objects, each of which may
have attributes of its own as well as relationships to other objects.

4
 Constraint Satisfaction Problems (CSPs)
• Standard search problem:
– state is a “black box” – any data structure that supports successor function,
heuristic function, and goal test

• CSP:
– state is defined by variables Xi (i=1,…,n) with values from domain Di
– goal test is a set of constraints Ci (i=1,…,m) specifying allowable
combinations of values for subsets of variables

• Simple example of a formal representation language

• Allows useful general-purpose algorithms with more power

than standard search algorithms

5
 Example: Map-Coloring

• Variables WA, NT, Q, NSW, V, SA, T

• Domains Di = {red, green, blue}

• Constraints: adjacent regions must have different colors

– e.g., WA ≠ NT, or (WA,NT) in {(red ,green), (red, blue), (green, red), (green, blue), (blue,
red), (blue, green)}

6
 Example: Map-Coloring

Solutions are complete and consistent assignments, e.g.,

WA = red, NT = green,Q = red,NSW = green,V = red,SA
= blue,T = green

7
 Constraint graph
• Binary CSP: each constraint relates two variables

• Constraint graph: nodes are variables, arcs are constraints

8
 CSP: Incremental Formulation
• Initial state: the empty assignment {}, in which all variables are
unassigned.
• Successor function: a value can be assigned to any unassigned
variable, provided that it does not conflict with previously
assigned variables.
• Goal test: the current assignment is complete.
• Path cost: a constant cost (e.g., 1) for every step.

9
 Example: N-Queens
• Formulation 1:
– Variables:
– Domains:
– Constraints

10
 Example: N-Queens

• Formulation 2:
– Variables:

– Domains:

– Constraints: Implicit:
Explicit:

11
 Example: Cryptarithmetic

• Variables: F T U W R O X1 X2 X3

• Domains: {0,1,2,3,4,5,6,7,8,9}

• Constraints: Alldiff (F,T,U,W,R,O)

– O + O = R + 10 ·X1
– X1 + W + W = U + 10 ·X2
– X2 + T + T = O + 10 ·X3
– X3 = F, T ≠ 0, F ≠ 0

12
 Example: Sudoku

• Variables: Each (open)

square
• Domains: {1,2,…,9}
• Constraints:
9-way alldiff for each column
9-way alldiff for each row
9-way alldiff for each region
(or can have a bunch
of pairwise inequality
constraints)

13
 Varieties of CSPs
• Discrete variables
– finite domains:
• n variables, domain size d → O(dn) complete assignments
– infinite domains:
• integers, strings, etc.
• e.g., job scheduling, variables are start/end days for each job
• need a constraint language, e.g., StartJob1 + 5 ≤ StartJob3

• Continuous variables
– e.g., start/end times for Hubble Space Telescope observations
– linear constraints solvable in polynomial time by linear
programming

14
 Varieties of constraints
• Unary constraints involve a single variable,
– e.g., SA ≠ green

• Binary constraints involve pairs of variables,

– e.g., SA ≠ WA

• Higher-order constraints involve 3 or more variables,

– e.g., cryptarithmetic column constraints

15
 Real-world CSPs
• Assignment problems
– e.g., who teaches what class

• Timetabling problems
– e.g., which class is offered when and where?

• Transportation scheduling

• Factory scheduling

• Notice that many real-world problems involve real-

valued variables

16
 Standard search formulation (incremental)

States are defined by the values assigned so far

• Initial state: the empty assignment { }
• Successor function: assign a value to an unassigned variable that
does not conflict with current assignment
– → fail if no legal assignments
• Goal test: the current assignment is complete

1. This is the same for all CSPs

2. Every solution appears at depth n with n variables
→ use depth-first search
3. Path is irrelevant, so can also use complete-state formulation
4. b = (n - l )d at depth l, hence n! ·dn leaves
17
 Backtracking search
• Depth-first search for CSPs with single-variable assignments is
called backtracking search

• Backtracking search is the basic uninformed algorithm for

CSPs

• Only need to consider assignments to a single variable at each

node ( b = d )
– There are n! ·dn leaves, dn possible complete assignments

18
 Backtracking Search
• Backtracking search = DFS + two improvements

• Idea 1: One variable at a time

– Variable assignments are commutative
• [WA = red then NT = green] same as [NT = green then
WA = red]
– Only need to consider assign value to a single variable at each
step

• Idea 2: Check constraints as you go

– Consider only values which do not conflict previous
assignments
– May need some computation to check the constraints
– “Incremental goal test”

19
 Backtracking Search

https://inst.eecs.berkeley.edu/~cs188/fa19/assets/demos/csp/csp_demos.html

20
 Backtracking Search

No need to check consistency for a complete assignment

21
 Backtracking Search

Checks consistency at each assignment

22
 Backtracking Search

• Backtracking = DFS + variable-ordering + fail-on-violation

23
 Backtracking example

24
 Backtracking example

25
 Backtracking example

26
 Backtracking example

27
 Improving backtracking efficiency
• Consider the order of variables to be assigned
and their values to be tried
– Which variable should be assigned next?
– In what order should its values be tried?
– Can we detect inevitable failure early?

• Interleaving search and inference

28
 Which variable should be
assigned next？

29
 Which variable should be assigned next？
• Most constrained variable:
choose the variable with the fewest legal values
– a.k.a. minimum remaining values (MRV) heuristic
• Most constraining variable (Tie-breaker):
– choose the variable with the most constraints on
remaining variables
– the degree heuristic

30
 Least constraining value
• Given a variable, choose the least constraining
value:
– The one that rules out the fewest values in the
remaining variables

31
 Interleaving search and inference
• Inference in CSPs
• In regular state-space search, algorithms only
do search. In CSPs, algorithms can search and
inference
– Search: choose a new variable assignment from
several possibilities
– Inference (a specific type): constraint propagation
• Constraint propagation may be intertwined
with search, or it may be done as a
preprocessing step, before search starts.
32
 Forward checking
(Interleaving search and inference)
• Idea:
– Keep track of remaining legal values for unassigned
variables
– Terminate search when any variable has no legal values

NT
Q
WA
SA
NSW

V
T

33
 Forward checking
• Idea:
– Keep track of remaining legal values for unassigned
variables
– Terminate search when any variable has no legal values

34
 Forward checking
• Idea:
– Keep track of remaining legal values for unassigned
variables
– Terminate search when any variable has no legal values

35
 Forward checking
• Idea:
– Keep track of remaining legal values for unassigned
variables
– Terminate search when any variable has no legal values

36
 Constraint propagation
• Forward checking propagates information from assigned to
unassigned variables, but doesn't provide early detection for all
failures:

• NT and SA cannot both be blue!

• Constraint propagation repeatedly enforces constraints locally

37
 Limitation of forward checking

• Although forward checking detects many

inconsistencies, it does not detect all of them.
– The problem is that it makes the current variable
arc-consistent, but doesn’t look ahead and make all
the other variables arc-consistent.

38
 Arc consistency
• Simplest form of propagation: making each arc consistent
• X →Y is consistent iff
for every value x of X there is some allowed y

39
 Consistency of A Single Arc

(Remove values from the tail!)

Recall: Binary constraint graph for a binary CSP (i.e., each

constraint has most two variables): nodes are variables, edges
show constraints

40
 Consistency of A Single Arc

41
 Arc consistency
• Simplest form of propagation: making each arc consistent
• X →Y is consistent iff
for every value x of X there is some allowed y

NSW → SA

42
 Arc consistency
• Simplest form of propagation: making each arc consistent
• X →Y is consistent iff
for every value x of X there is some allowed y

• If X loses a value, neighbors of X need to be rechecked

V → NSW

43
 Arc consistency
• Simplest form of propagation: making each arc consistent
• X →Y is consistent iff
for every value x of X there is some allowed y
• If X loses a value, neighbors of X need to be rechecked
• Arc consistency detects failure earlier than forward checking
• Can be run as a preprocessor or after each assignment

SA → NT

44
 How to Enforce Arc Consistency
of Entire CSP
• A simplistic algorithm: Cycle over the pairs of variables,
enforcing arc-consistency, repeating the cycle until no domains
change for a whole cycle
• AC-3 (short for Arc Consistency Algorithm #3): A more efficient
algorithm ignoring constraints that have not been modified since
they were last analyzed

NT
Q
WA
SA
NSW

V
T

45
 AC-3: Enforce Arc Consistency
of Entire CSP

Constraint Propagation!

46
 AC-3: Enforce Arc Consistency
of Entire CSP

Queue:
NT
Q
SA->WA
WA
SA
NT->WA
NSW

V
T

Remember: Delete from the tail!

47
 AC-3: Enforce Arc Consistency of
Entire CSP

Queue:
NT
Q
NT->WA
WA
SA
WA->SA
NSW NT->SA
V Q->SA
T
NSW->SA
V->SA

Remember: Delete from the tail!

48
 AC-3: Enforce Arc Consistency of
Entire CSP

Queue:
NT
Q
WA->SA
WA
SA
NT->SA
NSW Q->SA
V NSW->SA
T
V->SA
WA->NT
SA->NT
Q->NT

Remember: Delete from the tail!

49
 AC-3: Enforce Arc Consistency of
Entire CSP

Queue:
NT
Q
WA->SA
WA
SA
NT->SA
NSW Q->SA
V NSW->SA
T
V->SA
WA->NT
SA->NT
Q->NT

Remember: Delete from the tail!

50
 AC-3: Enforce Arc Consistency of
Entire CSP

Queue:
NT
Q
NT->SA
WA
SA
Q->SA
NSW NSW->SA
V V->SA
T
WA->NT
SA->NT
Q->NT

Remember: Delete from the tail!

51
 AC-3: Enforce Arc Consistency of
Entire CSP

Queue:
NT
Q
Q->SA
WA
SA
NSW->SA
NSW V->SA
V WA->NT
T
SA->NT
Q->NT

Remember: Delete from the tail!

52
 AC-3: Enforce Arc Consistency of
Entire CSP

Queue:
NT
Q
NSW->SA
WA
SA
V->SA
NSW WA->NT
V SA->NT
T
Q->NT

Remember: Delete from the tail!

53
 AC-3: Enforce Arc Consistency of
Entire CSP

Queue:
NT
Q
WA
SA
NSW

V
T

Remember: Delete from the tail!

54
 After assigning Q to Green,
what gets added to the Queue?

Queue:

NT
Q
WA
SA
NSW

V
T

A: NSW->Q, SA->Q, NT->Q

B: Q->NSW, Q->SA, Q->NT

55
 AC-3: Enforce Arc Consistency
of Entire CSP

Queue:
NT
Q
NT->Q
WA
SA
SA->Q
NSW NSW->Q
V
T

Remember: Delete from the tail!

56
 AC-3: Enforce Arc Consistency
of Entire CSP

Queue:
NT
Q
SA->Q
WA
SA
NSW->Q
NSW WA->NT
V SA->NT
T
Q->NT

Remember: Delete from the tail!

57
 AC-3: Enforce Arc Consistency
of Entire CSP

Queue:
NT
Q
NSW->Q
WA
SA
WA->NT
NSW SA->NT
V Q->NT
T
WA->SA
NT->SA
Q->SA
NSW->SA
V->SA

Remember: Delete from the tail!

58
 AC-3: Enforce Arc Consistency
of Entire CSP
Queue:
WA->NT
NT
SA->NT
WA
Q Q->NT
SA
NSW WA->SA
V NT->SA
T Q->SA
NSW->SA
V->SA
V->NSW
Q->NSW
SA->NSW

Remember: Delete from the tail!

59
 AC-3: Enforce Arc Consistency
of Entire CSP
Queue:
WA->NT
NT
SA->NT
WA
Q Q->NT
SA
NSW WA->SA
V NT->SA
T Q->SA
NSW->SA
V->SA
V->NSW
Q->NSW
SA->NSW

Remember: Delete from the tail!

60
 AC-3: Enforce Arc Consistency
of Entire CSP

Queue:
SA->NT
NT
Q->NT
WA
Q WA->SA
SA
NSW NT->SA
V Q->SA
NSW->SA
!!!
T
V->SA
V->NSW
Q->NSW
SA->NSW

Remember: Delete from the tail!

61
 AC-3: Enforce Arc Consistency
of Entire CSP
Queue:
SA->NT
Q->NT
NT
Q
WA->SA
WA
SA
NT->SA
NSW
Q->SA
V
NSW->SA
V->SA
!!!
T

V->NSW
Q->NSW
SA->NSW
• Backtrack on the assignment of Q
• Arc consistency detects failure earlier than forward checking
• Can be run as a preprocessor or after each assignment
• What’s the downside of enforcing arc consistency?

62
 Limitations of Arc Consistency
• After enforcing arc consistency:
– Can have one solution left
– Can have multiple solutions left
– Can have no solutions left (and
not know it)
• Arc consistency only checks local
consistency conditions
• Arc consistency still runs inside a What went
wrong here?
backtracking search!

63
 Backtracking Search with AC-3

• Where do you run AC-3?

64
 Complexity of a single run of AC-3

Recall that the whole backtracking algorithm with AC-3 will call AC-3 many times

65
 Complexity of a single run of AC-3

• An arc is added after a removal

of value at a node
• 𝑛 node in total, each has ≤ 𝑑
values
• Total times of removal: 𝑂 𝑛𝑑

66
 Complexity of a single run of AC-3

• An arc is added after a removal

of value at a node
• 𝑛 node in total, each has ≤ 𝑑
values
• Total times of removal: 𝑂 𝑛𝑑
• After a removal, ≤ 𝑛 arcs added
• Total times of adding arcs:
𝑂(𝑛2 𝑑)

67
 Complexity of a single run of AC-3

• An arc is added after a removal

of value at a node
• 𝑛 node in total, each has ≤ 𝑑
values
• Total times of removal: 𝑂 𝑛𝑑
• After a removal, ≤ 𝑛 arcs added
• Total times of adding arcs: 𝑂(𝑛2 𝑑)

• Check arc consistency per arc:

𝑂(𝑑 2 )

Complexity of a single run of AC-3 is at most 𝑂(𝑛2 𝑑 3 )

68
 Local search for CSPs
• Hill-climbing, simulated annealing typically work
with "complete" states, i.e., all variables assigned
• To apply to CSPs:
– allow states with unsatisfied constraints
– operators reassign variable values
• Variable selection: randomly select any conflicted
variable
• Value selection by min-conflicts heuristic:
– choose value that violates the fewest constraints
– i.e., hill-climb with h(n) = total number of violated
constraints
69
 Example: 4-Queens
• States: 4 queens in 4 columns (44 = 256 states)
• Actions: move queen in column
• Goal test: no attacks
• Evaluation: h(n) = number of attacks

• Given random initial state, can solve n-queens in almost constant time for
arbitrary n with high probability (e.g., n = 10,000,000)

70
 Summary
• CSPs are a special kind of problem:

– states defined by values of a fixed set of variables

– goal test defined by constraints on variable values

• Backtracking = depth-first search with one variable assigned per node

• Variable ordering and value selection heuristics help significantly
• Forward checking prevents assignments that guarantee later failure
• Constraint propagation (e.g., arc consistency) does additional work to
constrain values and detect inconsistencies
• Iterative min-conflicts is usually effective in practice

Credit：
15-281, Artificial Intelligence: Representation and Problem Solving, CMU

71