ARTIFICIAL

INTELLIGENC
E
Review of Last Lecture
 5

Informed Search Methods

Heuristic = “to find”, “to discover”

• Blind search:
• Search the solution space in an uninformed manner
•They simply march through the search space until a solution is found.
• Costly with respect time, space or both
• Informed search.
• Search the solution space in informed manner -> HEURISTIC
SEARCH
• Don’t try to explore all possible search path, rather focus on
the path which get you closer to the goal state using some
kind of a
“ GUIDE”.
• A heuristic search makes use of available information in making the
search more efficient.
 6

HEURISTIC SEARCH

•Heuristics:
• Rules for choosing the branches in a state space that are
•most likely to lead to an acceptable problem solution.
•Used when:
• Computational costs are high E.g. Chess
•But, Heuristics are fallible. They can lead a search algorithm to a suboptimal
solution or fail to find a solution at all.
 7

Informed Search Methods

Heuristic = “to find”, “to discover”

• A heuristic is a rule or method that almost always improves the

decision process.
• Generally can’t be sure that you are near the goal, but
heuristic is a good guess for the purpose.
• E.g. mouse searching for cheese having fresh smell.
• Mouse use “fresh smell” as a guide to reach the goal state.
 7

Informed Search Methods

Heuristic = “to find”, “to discover”

For example, if in a shop with several checkouts, it is usually best to go to the

one with the shortest queue. This holds true in most cases but further information
could influence this –

(1) if you saw that there was a checkout with only one person in that queue but that

the person currently at the checkout had three trolleys full of shopping

(2) that at the fast- checkout all the people had only one item, you may choose to go

to the fast- checkout instead.

(3) you don't go to a checkout that doesn't have a cashier - it may have the shortest

queue but you'll never get anywhere

Heuristic

• Euclidian/Straight line Distance= Sqrt (X2-X1)^2+(Y2-Y1) ^2

• Manhattan distance = all possible moves ()
• No of Misplaced tiles
1 3 2 1 2 3
6 5 4 4 5 6
8 7 7 8
Heuristic search
• Straight line distance of each city form goal city..

A 5 B 3 C 2

S 6
G
3 3
D 4 E F
Heuristic search: road
map
The number show 3
the distance b/w two
3
cities
2
A B C

S 4 4
G
3
2
D 1 E 3 F
Best First Search

• A combination of depth first (DFS) and breadth first

search (BFS).

• DFS is good because a solution can be found

without computing all nodes and BFS is good
because it does not get trapped in dead ends.

• The Best First Search (BestFS) allows us to switch

between paths thus gaining the benefits of both
approaches.
BestFS1: Greedy Search
• This is one of the simplest BestFS strategy.

• Heuristic function: h(n) - prediction of path cost left to the goal.

• Greedy Search: “To minimize the estimated cost to reach

the
goal”.

• The node whose state is judged to be closest to the goal state

is always expanded first.

• Two route-finding methods (1) Straight line distance; (2) minimum

Manhattan Distance - movements constrained to horizontal and
vertical directions.
BestFS1: Greedy Search
A Route Plan – Arad to Bucharest
BestFS1: Greedy Search (cont)
BestFS1: Greedy Search (cont)
BestFS1: Greedy Search (cont)
• Noticed that the solution for A  S  F  B is not optimum. It is 32
miles longer than the optimal path A  S  R  P  B.
• The strategy prefers to take the biggest bite possible out of the
remaining cost to reach the goal, without worrying whether this is
the best in the long run - hence the name ‘greedy search’.
• Greed is one of the 7 deadly sins, but it turns out that GS perform
quite well though not always optimal.
BestFS1: Greedy Search (cont)
BestFS1: Greedy Search (cont)

• GS resembles DFS in the way it prefers to follow a single path all

• the way to the goal, but will back up when it hits a dead end.

• Thus suffering the same defects as DFS - not optimal

and is incomplete.

• The worst-case time complexity for GS is O(bm), where m is the

• max depth of the search space.

• With good heuristic function, the space and time complexity

can be reduced substantially.
 22

Best First Search Algorithm

•Algorithm: At any time, expand the most promising node
according to the heuristic
If the heuristic is always 0, best-first search is the same as
breadth-first search

Best-first search is a greedy method.

Greedy methods maximize short-term advantage without
worrying about long-term consequences.
 23

Best First Search

Method
Function best_First_Search;
Begin
open := [Start];
closed:=[];
while open != []
do begin
remove the
leftmost
state from
open, call it
X;
If X = goal
then return
the path
from Start to
X
Else
begin
Generat
e
children
of X;
 24

Best First Search Demonstration

A:10

B: C:
5 3

D:4 E: F:
2 6

G:0
Solution
 25

Best First Search Demonstration

1. Open [A:10] : closed []
2. Evaluate A:10; open [C:3,B:5]; closed [A:10]
3. Evaluate C:3; open [B:5,F:6]; closed
4. [C:3,A:10]
5. Evaluate B:5; open [E:2,D:4,F:6]; closed
[C:3,B:5,A:10].
Evaluate E:2; open [G:0,D:4,F:6];
6.
closed [E:2,C:3,B:5,A:10]
Evaluate G:0; the solution / goal is reached
 26

Best First Search Demonstration

 27

Best First Search Demonstration

1. Open [A5] : closed []
2. Evaluate A5; open [B4,C4,D6]; closed [A5]
3. Evaluate B4; open [C4,E5,F5,D6]; closed
4. [B4,A5]
5. Evaluate C4; open [H3,G4,E5,F5,D6]; closed [C4,B4,A5].
6. Evaluate H3; open [O2,P3,G4,E5,F5,D6]; closed [H3,C4,B4,A5]
7. Evaluate O2; open [P3,G4,E5,F5,D6]; closed [O2,H3,C4,B4,A5]
Evaluate P3; THE SOLUTION IS FOUND!
Simple heuristic search procedure:
Hill Climbing
• Named so because it resembles an eager, but blind
mountain climber, you move in such direction in which
your height increased.
• Used a heuristic function which measure your height, or
measure remaining distance.

• As it keeps no history, the algorithm cannot recover

from failures of its strategy
Hill
Climbing
• Simplest way to implement heuristic search
• Expand current state in the search and evaluate its
children
• The best child is selected for further expansion; neither
its siblings nor parents are retained
 12

Hill Climbing:
If the Node is better in terms of assigned heuristics, only then you proceed
to
that Node
Algorithm:
1. Start with current-state (cs) = initial state
2. Until cs = goal-state or there is no change in the cs
do:

(a) Get the successor of cs and use the EVALUATION

FUNCTION to assign a score to each successor
(b) If one of the successor has a better score than cs
then
set the new state to be the successor with the best
score.
Hill-climbing (cont)

•This simple policy has three

(a)
well- known drawbacks:
1. Local Maxima: a local
maximum
• as opposed to global (b)

maximum.
2. Plateaus: An area of the (c)
search
• space where evaluation
function
•is Figure 5.9 Local maxima, Plateaus and
ridge situation for Hill
• flat, thus requiring random Climbing

walk.
Optimal searches
• We have covered informed and uninformed
searches.
• Bothe have some advantages and
disadvantages.
• One common lacks in both:
• Immediately stop when find a solution
• Don’t look for an optimal solution
• Optimal search:
• Find all the possible solution using
either blind/heuristic
search.
• Choose the most optimal amongst
 Branch and Bound (B& B)
• Some how related to heuristic search.
• Cut of the unnecessary search.

3 A 6 E 1

2 4
S D G
4 3
B
2

8 5
C
Branch and Bound

3 4 8
A B C
11
5 B 10 G
9
E D
7
6
8 D D
A

D D D E E G C 9
10 12 12 12 11 11
Branch and bound
• Some how related to heuristic search.
• Cut of the unnecessary search.

D
A
E
B 9
9 F G
30
. Branch and bound
The no show the distance b/w two cities
•.
3
2
A B
3
C
S 4 4
G
3
2
D 1 E 3 F
A
*•
A* (pronounced "A star") is a best-first, graph
search algorithm that finds the least-cost path
from a given initial node to one goal node (out
of one or more possible goals).
• Improve version of branch and bound.
• Uses a distance-plus-cost heuristic function
(usually denoted f(x)) to determine the order in
which the search visits nodes in the tree.
• Initially called algorithm A. Since using this
algorithm yields optimal behavior for a given
heuristic, it has been called A*.
Distance-Plus-Cost Heuristics function
• The distance-plus-cost heuristic is a sum of two functions:
• The path-cost function (usually denoted g(x)): cost from the starting
node to the current node.
• A heuristic estimate of the distance to the goal (usually denoted
h(x)).
• Total Cost: f(x) = g(x) + h(x)
A* Search
Algorithm
• At every step, take node n from the front of the queue
• Enqueue all its successors n" with priorities:
• f(n") = g(n") + h(n")
= cost of getting to n" + estimated cost from n" to goal
• Terminate when a goal state is popped from the queue.
 A* Search
. The Algorithm
no show the distance b/w two cities
•. And the color no shows the under estimated
distance between that particular node and
the goal 2
node. 3
3 3
A B C 5
2
8
S 4 4
0 G
3 5 2
2
6 D E F
 42

Examples of
Heuristics
 43

Last Solution