Search Problems

(Where we see problems as the

exploration of alternatives)

R&N: Chap. 3, Sect. 3.1–2 + 3.6

1
Example: 8-Puzzle

8 2 1 2 3
3 4 7 4 5 6

5 1 6 7 8

Initial state Goal state

Search is about the

exploration of alternatives
2
Exploratory search is an old idea:
The Labyrinth and the Ariadne Thread

According to Greek mythology, Theseus came to Crete to slay the

Minotaur, a monster who lived in a Labyrinth. Ariadne gave Theseus a
ball of yarn which he unwound as he entered the Labyrinth. After
killing the Minotaur, Theseus traced the thread back to the entrance
of the Labyrinth, rejoined Ariadne, and successfully escaped Crete

3
Since the dawn of civilization, puzzles and
games that require the exploration of
alternative paths have fascinated mankind
and have been considered a challenge for
human intelligence

 Chess originated in Persia and India

about 4000 years ago
 Checkers appeared as early as 1600 B.C
in Egyptian paintings
 Go originated in China over 3000 years
ago
4
5
15-Puzzle
Introduced in 1878 by Sam Loyd, who dubbed
himself “America’s greatest puzzle-expert”

1 2 3 4

5 6 7 8

9 10 11 12

13 14 15
6
7
15-Puzzle
Sam Loyd offered $1,000 of his own money to
the first person who would solve the following
problem:

1 2 3 4 1 2 3 4

5 6 7 8 ? 5 6 7 8

9 10 11 12 9 10 11 12

13 14 15 13 15 14
8
But no one ever won the prize !!
9
8-Puzzle: State Space

...
8 2 7
3 4
8 2 5 1 6

3 4 7
5 1 6 8 2 8 2
3 4 7 3 4 7

5 1 6 5 1 6

10
8-Puzzle: Successor Function
8 2 7
3 4
5 1 6

8 2 8 2 7 8 2 7
3 4 7 3 4 6 3 4
5 1 6 5 1 5 1 6

11
Stating a Problem as
a Search Problem

S  State space S
1  Successor function:
3 2
x  S  SUCCESSORS(x)  2S
 Arc cost
 Initial state s0
 Goal test:
xS  GOAL?(x) =T or F

12
State Graph
 It is defined as follows:
• Each state is represented by a distinct
node
• An arc connects a node s to a node s’ if
s’  SUCCESSORS(s)
 The state graph may contain more than
one connected component

13
14
Solution to the Search Problem
 A solution is a path connecting the initial
node to a goal node (any one)

15
16
Solution to the Search Problem
 A solution is a path connecting the initial
to a goal node (any one)
 The cost of a path is the sum of the
edge costs along this path
 An optimal solution is a solution path of
minimum cost
 There might be no solution !

17
18
How big is the state space of
the (n2-1)-puzzle?

 8-puzzle  9! = 362,880 states

 15-puzzle  16! ~ 1.3 x 1012 states
 24-puzzle  25! ~ 1025 states

But only half of these states are

reachable from any given state

19
 Permutation Inversions
 Wlg, let the goal be: 1 2 3 4
5 6 7 8
9 10 11 12
13 14 15

 Let ni be the number of tiles j < i that appear after tile i

(from left to right and top to bottom)
 N = n2 + n3 +  + n15 + row number of empty tile
1 2 3 4 n2 = 0 n3 = 0 n4 = 0
5 10 7 8 n5 = 0 n6 = 0 n7 = 1
n8 = 1 n9 = 1 n10 = 4 N=7+4
9 6 11 12
n11 = 0 n12 = 0 n13 = 0
13 14 15 n14 = 0 n15 = 0 20
 Proposition: (N mod 2) is invariant under
any legal move of the empty tile
 Proof:
• Any horizontal move of the empty tile
leaves N unchanged
• A vertical move of the empty tile changes
N by an even increment

1 2 3 4 1 2 3 4
5 6 7 5 6 11 7
s= s’ = N(s’) = N(s) + 3 - 1
9 10 11 8 9 10 8
13 14 15 12 13 14 15 12
21
 Proposition: (N mod 2) is invariant under
any legal move of the empty tile

  For a goal state g to be reachable

from a state s, a necessary condition is
that N(g) and N(s) have the same parity

 It can be shown that this is also a

sufficient condition

  The state graph consists of two

connected components of equal size
22
 15-Puzzle
Sam Loyd offered $1,000 of his own money to
the first person who would solve the following
problem:

1 2 3 4 1 2 3 4

5 6 7 8 ? 5 6 7 8

9 10 11 12 9 10 11 12

13 14 15 13 15 14

N=4 N=5
So, the second state is
not reachable from the
first, and Sam Loyd took
no risk with his money ...

23
What is the Actual State Space?
a) The set of all states?
[e.g., a set of 16! states for the 15-puzzle]
b) The set of all states from which a given goal
state is reachable?
[e.g., a set of 16!/2 states for the 15-puzzle]
c) The set of all states reachable from a given
initial state?
In general, the answer is a)

24
What is the Actual State Space?
a) The set of all states?
[e.g., a set of 16! states for the 15-puzzle]
b) The set of all states from which a given goal
state is reachable?
[e.g., a set of 16!/2 states for the 15-puzzle]
c) The set of all states reachable from a given
initial state?
In general, the answer is a)
But a fast test determining whether a state is reachable
from another is very useful, as search-based problem
solvers are often very inefficient when a problem has no
solution
More on this in future lectures ...
25
Stating a Problem as
a Search Problem

S  State space S
1  Successor function:
3 2
x  S  SUCCESSORS(x)  2S
 Arc cost
 Initial state s0
 Goal test:
xS  GOAL?(x) =T or F
 A solution is a path joining
the initial to a goal node
26
Searching the State Space
 Often it is not
feasible to build
a complete
representation
of the state
graph

27
8-, 15-, 24-Puzzles
8-puzzle  362,880 states

0.036 sec

15-puzzle  1.3 x 1012 states

< 4 hours

24-puzzle  1025 states

> 109 years

100 millions states/sec

28
Searching the State Space
 Often it is not
feasible to build
a complete
representation
of the state
graph
 A problem solver
must construct a
solution by
exploring a small
portion of the
graph

29
Searching the State Space

30
Searching the State Space

Search tree

31
Searching the State Space

Search tree

32
Searching the State Space

Search tree

33
Searching the State Space

Search tree

34
Searching the State Space

Search tree

35
Simple Problem-Solving-Agent
Algorithm

1. s0  sense/read initial state

2. GOAL?  select/read goal test
3. Succ  select/read successor function
4. solution  search(s0, GOAL?, Succ)
5. perform(solution)

36
 State Space
 Each state is an abstract representation of a
collection of possible worlds sharing some crucial
properties and differing on non-important details
only
E.g.: In assembly planning, a state does not define exactly
the absolute position of each part

 The state space is discrete. It may be finite, or

infinite

37
Successor Function
 It implicitly represents all the actions
that are feasible in each state

38
Successor Function
 It implicitly represents all the actions
that are feasible in each state
 Only the results of the actions (the
successor states) and their costs are
returned by the function
 The successor function is a “black box”:
its content is unknown
E.g., in assembly planning, the function does
not say if it only allows two sub-assemblies to
be merged or if it makes assumptions about
subassembly stability 39
Path Cost
 An arc cost is a positive number
measuring the “cost” of performing the
action corresponding to the arc, e.g.:
• 1 in the 8-puzzle example
• expected time to merge two sub-assemblies
 We will assume that for any given
problem the cost c of an arc always
verifies: c  ε  0, where ε is a constant

40
Path Cost
 An arc cost is a positive number
measuring the “cost” of performing the
action corresponding to the arc, e.g.:
• 1 in the 8-puzzle example
• expected time to merge two sub-assemblies
 We will assume that for any given
problem the cost c of an arc always
verifies: c  ε  0, where ε is a constant
[This condition guarantees that, if path becomes
arbitrarily long, its cost also becomes arbitrarily large]
Why is this needed? 41
Goal State 1 2 3
 It may be explicitly described: 4 5 6
7 8
1 a a
 or partially described: a
5 a
(“a” stands for “any”)
a 8 a

 or defined by a condition,
e.g., the sum of every row, of every column,
and of every diagonals equals 30 15 1 2 12
4 10 9 7
8 6 5 11
3 13 14
42
Other examples

43
8-Queens Problem
Place 8 queens in a chessboard so that no two
queens are in the same row, column, or diagonal.

A solution Not a solution

44
Formulation #1
 States: all arrangements of 0,
1, 2, ..., or 8 queens on the
board
 Initial state: 0 queen on the
board
 Successor function: each of
the successors is obtained by
adding one queen in an empty
square
 Arc cost: irrelevant
 Goal test: 8 queens are on the
board, with no two of them
attacking each other
 64x63x...x53 ~ 3x1014 states
45
Formulation #2
 States: all arrangements of k = 0,
1, 2, ..., or 8 queens in the k
leftmost columns with no two
queens attacking each other
 Initial state: 0 queen on the board
 Successor function: each successor
is obtained by adding one queen in
any square that is not attacked by
any queen already in the board, in
the leftmost empty column
 Arc cost: irrelevant
 Goal test: 8 queens are on the
board

 2,057 states
46
n-Queens Problem
 A solution is a goal node, not a path to this
node (typical of design problem)
 Number of states in state space:
• 8-queens  2,057
• 100-queens  1052
 But techniques exist to solve n-queens
problems efficiently for large values of n
They exploit the fact that there are many
solutions well distributed in the state space

47
Path Planning

What is the state space?

48
Formulation #1

Cost of one horizontal/vertical step = 1

Cost of one diagonal step = 2
49
Optimal Solution

This path is the shortest in the discretized state

space, but not in the original continuous space
50
 Formulation
sweep-line
#2

51
Formulation #2

52
States

53
Successor Function

54
Solution Path

A path-smoothing post-processing step is

usually needed to shorten the path further
55
Formulation #3

Cost of one step: length of segment

56
Formulation #3

Visibility graph

Cost of one step: length of segment

57
Solution Path

The shortest path in this state space is also the

shortest in the original continuous space
58
Assembly (Sequence) Planning

59
Possible Formulation
 States: All decompositions of the assembly
into subassemblies (subsets of parts in their
relative placements in the assembly)
 Initial state: All subassemblies are made of a
single part
 Goal state: Un-decomposed assembly
 Successor function: Each successor of a state
is obtained by merging two subassemblies (the
successor function must check if the merging
is feasible: collision, stability, grasping, ...)
 Arc cost: 1 or time to carry the merging
60
A Portion of State Space

61
But the formulation rules out
“non-monotonic” assemblies

62
But the formulation rules out
“non-monotonic” assemblies

63
But the formulation rules out
“non-monotonic” assemblies

64
But the formulation rules out
“non-monotonic” assemblies

65
But the formulation rules out
“non-monotonic” assemblies

66
But the formulation rules out
“non-monotonic” assemblies

X
This “subassembly” is not
allowed in the definition of
the state space: the 2 parts
are not in their relative
placements in the assembly

Allowing any grouping of parts

as a valid subassembly would
make the state space much
bigger and more difficult
to search
67
Assumptions in Basic Search

 The world is static

 The world is discretizable
 The world is observable
 The actions are deterministic

But many of these assumptions can be

removed, and search still remains an
important problem-solving tool

68
Vacuum Cleaner Problem
 A vacuum robot lives in a two-room environment

 States: The robot is in one of the two rooms,

and each room may or may not contain dirt
 8 states
 Successor function: the successors of a state
correspond to trying 3 actions: Right, Left,
Suck.
 Initial state: Unknown (not observable)
 Goal state: No dust in the rooms
69
Re-Formulation with “Belief States”
 Belief states: sets of states  28 =256
belief states
 Initial belief state: set of 8 states
 Successor function: the successors of a
belief state correspond to trying Right,
Left, Suck on all states in the belief
state.
 Goal belief state: any set of states with
none having dust in any of the rooms
70
Left Suck Right

71
Search and AI
 Search methods are ubiquitous in AI systems.
They often are the backbones of both core
and peripheral modules
 An autonomous robot uses search methods:
• to decide which actions to take and which sensing
operations to perform,
• to quickly anticipate and prevent collision,
• to plan trajectories,
• to interpret large numerical datasets provided by
sensors into compact symbolic representations,
• to diagnose why something did not happen as
expected,
• etc...
72
Applications
Search plays a key role in many applications, e.g.:

 Route finding: airline travel, networks

 Package/mail distribution
 Pipe routing, VLSI routing
 Comparison and classification of protein folds
 Pharmaceutical drug design
 Design of protein-like molecules
 Inverse analysis for non-destructive testing
 Video games
73