Heuristic Search

Chapter 3
 Outline

Generate-and-test
Hill climbing
Best-first search
Problem reduction
Constraint satisfaction
Means-ends analysis

2
 Generate-and-Test

Algorithm
1. Generate a possible solution.
2. Test to see if this is actually a solution.
3. Quit if a solution has been found.
Otherwise, return to step 1.

3
 Generate-and-Test

Acceptable for simple problems.

Inefficient for problems with large space.

4
 Generate-and-Test

Exhaustive generate-and-test.
Heuristic generate-and-test: not consider paths that
seem unlikely to lead to a solution.
Plan generate-test:
- Create a list of candidates.
- Apply generate-and-test to that list.

5
 Generate-and-Test

Example: coloured blocks

Arrange four 6-sided cubes in a row, with each side of
each cube painted one of four colours, such that on all four
sides of the row one block face of each colour is showing.

6
 Generate-and-Test

Example: coloured blocks

Heuristic: if there are more red faces than other colours

then, when placing a block with several red faces, use few
of them as possible as outside faces.

7
 Hill Climbing

Searching for a goal state = Climbing to the top of a hill

8
 Hill Climbing

Generate-and-test + direction to move.

Heuristic function to estimate how close a given state
is to a goal state.

9
 Simple Hill Climbing

Algorithm
1. Evaluate the initial state.
2. Loop until a solution is found or there are no new
operators left to be applied:
- Select and apply a new operator
- Evaluate the new state:
goal quit
better than current state new current state

10
 Simple Hill Climbing

Evaluation function as a way to inject task-specific

knowledge into the control process.

11
 Simple Hill Climbing

Example: coloured blocks

Heuristic function: the sum of the number of different

colours on each of the four sides (solution = 16).

12
 Steepest-Ascent Hill Climbing
(Gradient Search)
Considers all the moves from the current state.
Selects the best one as the next state.

13
 Steepest-Ascent Hill Climbing
(Gradient Search)
Algorithm
1. Evaluate the initial state.
2. Loop until a solution is found or a complete iteration
produces no change to current state:
- SUCC = a state such that any possible successor of the
current state will be better than SUCC (the worst state).
- For each operator that applies to the current state, evaluate
the new state:
goal quit
better than SUCC set SUCC to this state
- SUCC is better than the current state set the current
state to SUCC. 14
 Hill Climbing: Disadvantages

Local maximum
A state that is better than all of its neighbours, but not
better than some other states far away.

15
 Hill Climbing: Disadvantages

Plateau
A flat area of the search space in which all neighbouring
states have the same value.

16
 Hill Climbing: Disadvantages

Ridge
The orientation of the high region, compared to the set
of available moves, makes it impossible to climb up.
However, two moves executed serially may increase
the height.

17
 Hill Climbing: Disadvantages

Ways Out

Backtrack to some earlier node and try going in a

different direction.
Make a big jump to try to get in a new section.
Moving in several directions at once.

18
 Hill Climbing: Disadvantages

Hill climbing is a local method:

Decides what to do next by looking only at the
immediate consequences of its choices.
Global information might be encoded in heuristic
functions.

19
Hill Climbing: Disadvantages

Start Goal
A D
D C
C B
B A
Blocks World

20
 Hill Climbing: Disadvantages

Start Goal
A D
0 4
D C
C B
B A
Blocks World
Local heuristic:
+1 for each block that is resting on the thing it is supposed to be
resting on.
-1 for each block that is resting on a wrong thing.
21
Hill Climbing: Disadvantages

0 A 2

D D
C C
B B A

22
 Hill Climbing: Disadvantages

D 2
C
B A

A 0
0
D
C C D C 0

B B A B A D

23
 Hill Climbing: Disadvantages

Start Goal
A D
-6 6
D C
C B
B A
Blocks World
Global heuristic:
For each block that has the correct support structure: +1 to
every block in the support structure.
For each block that has a wrong support structure: -1 to
every block in the support structure. 24
 Hill Climbing: Disadvantages

D -3
C
B A

A -6
-2
D
C C D C -1

B B A B A D

25
 Hill Climbing: Conclusion

Can be very inefficient in a large, rough problem space.

Global heuristic may have to pay for computational

complexity.

Often useful when combined with other methods, getting

it started right in the right general neighbourhood.

26
 SimulatedAnnealing

Avariation of hill climbing in which, at the beginning of

the process, some downhill moves may be made.

To do enough exploration of the whole space early on, so

that the final solution is relatively insensitive to the
starting state.
Lowering the chances of getting caught at a local
maximum, or plateau, or a ridge.

27
 SimulatedAnnealing

PhysicalAnnealing
Physical substances are melted and then gradually
cooled until some solid state is reached.
The goal is to produce a minimal-energy state.
Annealing schedule: if the temperature is lowered
sufficiently slowly,then the goal will be attained.
Nevertheless, there is some probability for a transition
to a higher energy state: e -E/kT.

28
 SimulatedAnnealing

Algorithm
1. Evaluate the initial state.
2. Loop until a solution is found or there are no new
operators left to be applied:
- Set T according to an annealing schedule
- Selects and applies a new operator
- Evaluate the new state:
goal quit
E = V
a l(current state) - V
a l(new state)
E < 0 new current state
else new current state with probability e -E/kT. 29
 Best-First Search

Depth-first search: not all competing branches having to

be expanded.

Breadth-first search: not getting trapped on dead-end

paths.
Combining the two is to follow a single path at a
time, but switch paths whenever some competing
path look more promising than the current one.

30
 Best-First Search

A A A

B C D B C D
3 5 1 3 5
E F
4 6
A A

B C D B C D
5 5
G H E F G H E F
6 5 4 6 6 5 6
I J
2 1
31
 Best-First Search

OPEN: nodes that have been generated, but have not

examined.
This is organized as a priority queue.

CLOSED: nodes that have already been examined.

Whenever a new node is generated, check whether it
has been generated before.

32
 Best-First Search

Algorithm
1. OPEN = {initial state}.
2. Loop until a goal is found or there are no nodes left in
OPEN:
- Pick the best node in OPEN
- Generate its successors
- For each successor:
new evaluate it, add it to OPEN, record its parent
generated before change parent, update successors

33
 Best-First Search

Greedy search:
h(n) = estimated cost of the cheapest path from node
n to a goal state.

34
 Best-First Search

Uniform-cost search:
g(n) = cost of the cheapest path from the initial state
to node n.

35
 Best-First Search

Greedy search:
h(n) = estimated cost of the cheapest path from node
n to a goal state.
Neither optimal nor complete

36
 Best-First Search

Greedy search:
h(n) = estimated cost of the cheapest path from node
n to a goal state.
Neither optimal nor complete

Uniform-cost search:
g(n) = cost of the cheapest path from the initial state
to node n.
Optimal and complete, but very inefficient

37
 Best-First Search

Algorithm A* (Hart et al., 1968):

f(n) = g(n) + h(n)

h(n) = cost of the cheapest path from node n to a

goal state.
g(n) = cost of the cheapest path from the initial state
to node n.

38
 Best-First Search

Algorithm A*:

f*(n) = g*(n) + h*(n)

h*(n) (heuristic factor) = estimate of h(n).

g*(n) (depth factor) = approximation of g(n) found by

A* so far.

39
 Problem Reduction

Goal:Acquire TV set

Goal: Steal TV set Goal: Earn some money Goal: Buy TV set

AND-OR Graphs

Algorithm AO* (Martelli & Montanari 1973, Nilsson 1980)

40
 Problem Reduction:AO*

A A 6
5 9
B C D
3 4 5

A 9 A 1
9 12
B C D 10 B 6 C D 10
3 4 4
E F G H E F
4 4 5 7 4 4

41
 Problem Reduction:AO*

A 1 A 14

B 13 C 10 B 13 C 15

D 5 E 6 F 3 D 5 E 6 F 3

G 5 G 10

Necessary backward propagation H 9

42
 Constraint Satisfaction

ManyAI problems can be viewed as problems of

constraint satisfaction.

Cryptarithmetic puzzle:

SEND

MORE
MONEY

43
 Constraint Satisfaction

As compared with a straightforard search procedure,

viewing a problem as one of constraint satisfaction can
reduce substantially the amount of search.

44
 Constraint Satisfaction

Operates in a space of constraint sets.

Initial state contains the original constraints given in the

problem.

Agoal state is any state that has been constrained

enough.

45
 Constraint Satisfaction

Two-step process:
1. Constraints are discovered and propagated as far as
possible.
2. If there is still not a solution, then search begins,
adding new constraints.

46
Initial state:
No two letters have
the same value. M=1
The sum of the digits S = 8 or 9 SEND
must be as shown.
O=0
N=E+1 MORE
C2 = 1
N+R>8
E9 MONEY
E=2
N=3
R = 8 or 9
2 + D = Y or 2 + D = 10 + Y

C1 = 0 C1 = 1

2 + D =Y 2 + D = 10 + Y
N + R = 10 + E D = 8 +Y
R=9 D = 8 or 9
S =8
D=8 D=9
Y= 0 Y= 1
47
 Constraint Satisfaction

Two kinds of rules:

1. Rules that define valid constraint propagation.
2. Rules that suggest guesses when necessary.

48
 Homework

Exercises 1-14 (Chapter 3 AI Rich & Knight)

Reading Algorithm A*
([Link]

49