ARTIFICIAL

INTELLIGENCE- CS411
Prof. Alaa Sagheer
Artificial Intelligence “CS 411”
• Textbook:
S. Russell and P. Norvig
Artificial Intelligence: A Modern Approach
Prentice Hall, 2020, Fourth Edition

• Place: Male and Female Campus

• Grading:
Homework & Activities (5%),
Project @ Lab (10%),
Quizzes @ Class (10%),
Quizzes @ Lab (15 %),
Mid-term exam (20%),
Final exam (40%),
Artificial Intelligence Course Info Prof. Alaa Sagheer
 Problem Solving

Introduction
“Graph and Tree”

Lecture I S
 4

A Problem?

Artificial Intelligence Ch3: Problem Solving by Search Prof. Alaa Sagheer

 5

A Problem?

Artificial Intelligence Ch3: Problem Solving by Search Prof. Alaa Sagheer

 6

A Problem?

Example: 8 - Puzzles

Artificial Intelligence Ch3: Problem Solving by Search Prof. Alaa Sagheer

 7

Graphs VS. Trees

Each NODE in the search tree is an entire PATH

in the state graph (how many nodes in the graph
appear multiple times in the search tree)

Artificial Intelligence Ch3: Problem Solving by Search Prof. Alaa Sagheer

 8

Tree Search Algorithm

Basic idea: systematic exploration of simulated state-space by generating
successors of explored states

Artificial Intelligence Ch3: Problem Solving by Search Prof. Alaa Sagheer

 3- Problem Solving

Solving Problem by
Searching

Lecture I S
 Course Learning Outcomes
Chapter 3
Problem Solving
At the end of the course students will be able to
1. Compare different types of learning paradigms.
2. Apply suitable learning algorithm to a classification task.
3. Compare and contrast common models used for knowledge representation.

4. Design and implement an AI algorithm based

solution to a problem.
5. Compare and contrast various techniques for goal search.
6. Explain the distinction between various types of reasoning and inference techniques.
7. Apply rule-based, case-based and model-based reasoning techniques.

10

Artificial Intelligence Ch3: Problem Solving by Search Prof. Alaa Sagheer

 11

• Simplest agents (reflex agents, which base their actions on a direct mapping from
states to actions) cannot operate well in environments for which this mapping
would be too large to store and would take too long to learn.

• Goal-based agents can succeed by considering future actions and the desirability
of their outcomes.
Problem-Solving Agents

We have a tourist would

like to travel from Arad to
Bucharest. Of course he
will seek about which path
with achieves his goal and
go to required destination.
Artificial Intelligence Ch3: Problem Solving by Search Prof. Alaa Sagheer
 12

• Simplest agents (reflex agents, which base their actions on a direct mapping from
states to actions) cannot operate well in environments for which this mapping
would be too large to store and would take too long to learn.

• Goal-based agents can succeed by considering future actions and the desirability
of their outcomes.
Problem-Solving Agents
Are Deciding what to do by finding sequences of actions that lead to desirable states
•Intelligent agents are supposed to maximize their performance measure.
Achieving this is sometimes simplified if the agent can adopt a goal, (e.g. Taxi)
•Goals help to organize Agent’s behavior by limiting the objectives that the agent
is trying to achieve.
•Goal formulation, based on the current situation and the agent's performance
measure, is the first step in problem solving.
•A goal is a set of world states- exactly those states in which the goal is satisfied.
The agent's task is to find out which sequence of actions will get it to a goal state
*.
Artificial Intelligence Ch3: Problem Solving by Search Prof. Alaa Sagheer
 13

Problem-Solving by Search
An agent with several immediate options of unknown value can decide
what to do by just examining different possible sequences of actions that
lead to states of known value, and then choosing the best sequence.

This Process of looking for such a sequence is called Search

A search algorithm takes a problem as input and returns a

solution in the form of an action sequence, once a solution is
found, the actions it recommends can be carried out.

Formulate, Search, Execute

Artificial Intelligence Ch3: Problem Solving by Search Prof. Alaa Sagheer
 14

Problem-Solving by Search

After formulating a goal, the agent calls a search procedure to solve it. It then uses the
solution to guide its actions, doing whatever the solution recommends as the next thing
to do- typically, the first action of the sequence- and then removing that step from the
sequence. Once the solution has been executed, the agent will formulate a new goal.
Artificial Intelligence Ch3: Problem Solving by Search Prof. Alaa Sagheer
 15

Well-defined problem and solution (1)

A problem can be defined formally by four components:
1- The initial state that the agent starts in,
2- A description of the possible actions available to the agent
Successor Function
Given a particular state x, SUCCESSOR-FN(x) returns a set of (action, successor) ordered
pairs. where each action is one of the legal actions in state x and each successor is a state that
can be reached from x by applying the action.
Example: From the state In (Arad), the successor function for this problem would return:
{(Go(Sibiu), In (Sibiu)), (Go(Timisoara), In (Timisoara)), (Go(Zerind), In (Zerind))}
Together, the initial state and successor
function implicitly define the state
space of the problem, which is the
set of all states reachable from the
initial state. A path in the state
space is a sequence of states
connected by a sequence of
actions.
Artificial Intelligence Ch3: Problem Solving by Search Prof. Alaa Sagheer
 16

Well-defined problem and solution (2)

3- The goal test, which determines whether a given state is a goal state or
not. When we have explicit set of possible goal states, the test checks
whether the given state is one of them or not.

4- A path cost function that assigns a numeric cost to each path. This cost
function reflects the performance measure. (Example: Time to Bucharest
so the cost of a path might be its length). So the cost of a path might be
its length in kilometers.

A solution to a problem is a path from the initial state

to a destination state. Solution quality is measured by
the path cost function. An optimal solution has the
lowest path cost among all solutions.

Artificial Intelligence Ch3: Problem Solving by Search Prof. Alaa Sagheer

 17

Example of Problems
A toy problem is intended to illustrate (or exercise) various
problem-solving methods. It can be given a concise, exact
description. This means that it can be used easily by different
researchers to compare the performance of algorithms.
A real-world problem is one whose solutions people actually care
about. They tend not to have a single agreed-upon description

Toy Problems
❖ The Vacuum World
❖ The Puzzle World

Artificial Intelligence Ch3: Problem Solving by Search Prof. Alaa Sagheer

 18

Toy Problems (1)

❖ The Vacuum World
States: The agent is in one of two locations, each of which might or might
not contain dirt. Thus there are 2 x 22 = 8 possible world states. (n rooms??)
• Initial state: Any state can be designated as the initial state,
• Successor function: This generates the legal states that result from trying the
three actions (Left, Right, and Suck). The complete state space is shown below.

• Goal test: This checks whether all the

squares are clean.
• Path cost: Each step costs 1, so the path cost is
the number of steps in the path.
Compared with the real world, this toy problem has
discrete locations, discrete dirt, reliable cleaning, and it

never gets messed up once cleaned.

Artificial Intelligence Ch3: Problem Solving by Search Prof. Alaa Sagheer
 19

Toy Problems (2)

❖ The 8-Puzzle
It consists of a 3 x 3 board with eight numbered tiles and a blank space.
States: A state description specifies the location of each of the eight tiles and the
blank in one of the nine squares
• Initial state: Any state can be designated as the initial state. Note that any given
goal can be reached from exactly half of the possible initial states.
• Successor function: This generates the legal states that result from trying the four
actions (blank moves Left, Right, Up, or Down)
• Goal test: This checks whether the
state matches the goal configuration
(Other goal configurations are possible.)
• Path cost: Each step costs 1, so the
path cost is the number of steps in
the path.
The actions are abstracted to their beginning and final states, ignoring the intermediate locations
where the block is sliding.
Artificial Intelligence Ch3: Problem Solving by Search Prof. Alaa Sagheer
 20

Real-world Problems (1)

❖ The Airline travel problem
It is example of Route-finding problem (e.g. routing, military, TSP, …)
States: Each is represented by a location (e.g., an airport) and the current time.

• Initial state: This is specified by the problem.

• Successor function: This returns the states resulting from taking any scheduled
flight leaving later than the current time plus the within-airport transit time, from
the current airport to another
• Goal test: Are we at the destination by some prespecified time?
• Path cost: This depends on monetary cost, waiting time, flight time, customs and
immigration procedures, seat quality, time of day, type of airplane, frequent-flyer
mileage awards, and so on

Artificial Intelligence Ch3: Problem Solving by Search Prof. Alaa Sagheer

 21

Real-world Problems (2)

❖ The Travel Salesman Problem
It is example of Touring problem, in which each city must be visited
exactly once. The aim is to find the shortest tour (or path).
- The state space is quite different. Each state must include not just the current location
but also the set of cities the agent has visited.
- So the initial state would be
“In Bucharest; visited {Bucharest},”
- A typical intermediate state would be:
“In Vaslui; visited {Bucharest, Urziceni,
Vaslui},”

and the goal test would check whether

the agent is in Bucharest and all 20 cities
have been visited

Artificial Intelligence Ch3: Problem Solving by Search Prof. Alaa Sagheer

 22

Searching for Solutions (1)

Having formulated some problems, we now need to solve them. This is done by a search
through the state space. This chapter deals with search techniques that use an explilci search
tree that is generated by the initial state and the successor function that together define the
state space. The root of the search tree is a search node

Search node Search Tree

Artificial Intelligence Ch3: Problem Solving by Search Prof. Alaa Sagheer

 23

Searching for Solutions (2)

(a) The initial state

(b) After expanding Arad

(c) After expanding Sibiu

Artificial Intelligence Ch3: Problem Solving by Search Prof. Alaa Sagheer

 24

Searching for Solutions (3)

❖ Expanding Search Strategy

State space vs. Search tree ???

For the route finding problem, there are only 20 states in the state space. But there
are an infinite number of paths in this state space, so the search tree has an
infinite number of nodes.
For example, “Arad- Sibiu”, “Arad- Sibiu- Arad”, “Arad-Sibiu-Arad-Sibiu” are the
first three of an infinite sequence of paths, and so on.

Artificial Intelligence Ch3: Problem Solving by Search Prof. Alaa Sagheer

 25

Searching for Solutions (4)

❖ Expanding Search Strategy

Nodes vs. States ???

A state corresponds to a configuration of
the world.
A node is a data structure from which the
search tree is constructed. Each node has
a parent node, a state,
action, path cost g(x), depth.

Artificial Intelligence Ch3: Problem Solving by Search Prof. Alaa Sagheer

 26

Searching for Solutions (4)

❖ Expanding Search Strategy

The collection of nodes that have been

generated but not yet expanded is called
Fringe.
Each element of the fringe is a leaf node,
that is, a node with no successors in the
tree
Artificial Intelligence Ch3: Problem Solving by Search Prof. Alaa Sagheer
 27

Measuring problem-solving performance

The output of a problem-solving algorithm is either failure or a solution. We evaluate
an algorithm's performance in four ways:

▪ Completeness: Is the algorithm guaranteed to find a solution?

▪ Optimality: Does the strategy find the optimal solution?
▪ Time complexity: How long does it take to find a solution?
▪ Space complexity: How much memory is needed to perform the search?
- Time is often measured in terms of the number of nodes generated during the
search, and Space in terms of the maximum number of nodes stored in
memory.
- The effectiveness of a search algorithm depends on the search cost- which
typically depends on the time complexity and space complexity.
- Finally, we can use the total cost, which combines the search cost and the
path cost of the solution found.
Artificial Intelligence Ch3: Problem Solving by Search Prof. Alaa Sagheer
 28

Properties of Search strategies

Artificial Intelligence Ch3: Problem Solving by Search Prof. Alaa Sagheer

 29

Properties of Search strategies

Artificial Intelligence Ch3: Problem Solving by Search Prof. Alaa Sagheer

 30

Properties of Search strategies

Artificial Intelligence Ch3: Problem Solving by Search Prof. Alaa Sagheer

 31

Properties of Search strategies

Artificial Intelligence Ch3: Problem Solving by Search Prof. Alaa Sagheer

 32

Properties of Search strategies

• Time and space complexity are measured in terms of
• b: branching factor (i.e. maximum number of successors of any node)

• d: depth of the least-cost solution (or goal)

• m: maximum depth of any path in the state space (may be ∞)

Artificial Intelligence Ch3: Problem Solving by Search Prof. Alaa Sagheer

 3- Problem Solving

Uninformed
Search

Lecture II S
 34

Uninformed Search Strategies

• Uninformed (or Blind Search) search strategies use only the
information available in the problem definition..There is no
additional Information. All they can do is generate successors
and distinguish a goal state from a non goal state.
• Breadth-first search
•
• Uniform-cost search
•
• Depth-first search
•
• Depth-limited search
•
• Iterative deepening search
Artificial
• Intelligence Ch3: Problem Solving by Search Prof. Alaa Sagheer
 35

Tree Search
• A tree search starts at the root
and explores nodes from there,
A
looking for a goal node (a node
that satisfies certain conditions,
B C depending on the problem)

D E F G • For some problems, any goal

node is acceptable (N or J); for
other problems, you want a
H I J K minimum-depth goal node, that
is, a goal node nearest the root
L M N O P Q (only J)

Goal nodes

Artificial Intelligence Ch3: Problem Solving by Search Prof. Alaa Sagheer

 36

Breadth-First Search - Example

• A breadth-first search (BFS)

A explores nodes nearest the
root before exploring nodes
further away
B C
• For example, after searching
A, then B, then C, the search
D E F G proceeds with D, E, F, G
• Node are explored in the
H I J K order A B C D E F G H I J K L
MNOPQ

L M N O P Q • J will be found before N

Artificial Intelligence Ch3: Problem Solving by Search Prof. Alaa Sagheer

 37

Breadth-First Search
• The root node is expanded first, then all the successors of the
root node are expanded next, then their successor, and so on.
• Implementation:
• Fringe is a FIFO queue, i.e., new successors go at end. i.e. Nodes that
are visited first will be expanded first. FIFO = First-In-First-Out
•

Artificial Intelligence Ch3: Problem Solving by Search Prof. Alaa Sagheer

 38

Breadth-First Search
• The root node is expanded first, then all the successors of the
root node are expanded next, then their successor, and so on.
• Implementation:
• Fringe is a FIFO queue, i.e., new successors go at end. i.e. Nodes that
are visited first will be expanded first.
•

At each stage, the node to be expanded next is indicated by a red marker

Artificial Intelligence Ch3: Problem Solving by Search Prof. Alaa Sagheer

 39

Breadth-First Search
• The root node is expanded first, then all the successors of the
root node are expanded next, then their successor, and so on.
• Implementation:
• Fringe is a FIFO queue, i.e., new successors go at end. i.e. Nodes that
are visited first will be expanded first.
•

At each stage, the node to be expanded next is indicated by a red marker

Artificial Intelligence Ch3: Problem Solving by Search Prof. Alaa Sagheer

 40

Breadth-First Search
• The root node is expanded first, then all the successors of the
root node are expanded next, then their successor, and so on.
• Implementation:
• Fringe is a FIFO queue, i.e., new successors go at end. i.e. Nodes that
are visited first will be expanded first.
•

At each stage, the node to be expanded next is indicated by a red marker

Artificial Intelligence Ch3: Problem Solving by Search Prof. Alaa Sagheer

 41

Properties of breadth-first search

• Regarding Time and Memory, it is not always the strategy of
choice!
• Consider a state space where every state has b successors.
The root of the search tree generates b nodes at the first level,
each of which generates b more nodes, for a total of b2 at the
second level. Each of these generates b more nodes, yielding
b3 nodes at the third level, and so on.
• Now suppose that the solution is at depth d. In the worst case,
we would expand all but the last node at level d (since the goal
itself is not expanded), generating bd+1 - b nodes at level d+1*.

• Then the maximum total number of nodes generated is:

b + b2 + b3 +….+ bd + (bd+1- b) = O(bd+1)

Artificial Intelligence Ch3: Problem Solving by Search Prof. Alaa Sagheer

 42

Space is the bigger problem (more than time)

Artificial Intelligence Ch3: Problem Solving by Search Prof. Alaa Sagheer
 43

Cost-Sensitive Search
a GOAL
2 2
b c
3
2
1 8
2 e
3 d
f
9 8 2
START h
1 4 2

p 4 r
15
q

BFS finds the shortest path in terms of number of actions.

It does not find the least-cost path. We will now cover
a similar algorithm which does find the least-cost path.
Artificial Intelligence Ch3: Problem Solving by Search Prof. Alaa Sagheer
 44

Uniform-cost search
Breadth-first search is optimal when all step costs are equal,
because it always expands the shallowest unexpanded node.

• Instead of expanding the shallowest node, uniform-cost

search expands the node n with the lowest path cost in each
branch up to the Goal.

• Implementation:
• fringe = queue ordered by path cost
•
• Equivalent to breadth-first if step costs
• all equal
•

Artificial Intelligence Ch3: Problem Solving by Search Prof. Alaa Sagheer

 45

Uniform-cost search
2 a G
b c
Strategy: expand a cheapest node first: 1 8 2
2 e
Fringe is a priority queue (priority: cumulative cost) 3 d f
9 2
S h 8 1
1 p q r
15

S 0

d 3 e 9 p 1

b 4 c e 5 h 17 r 11 q 16
11
Cost a 6 a h 13 r 7 p q f
contours
p q f 8 q c G

q 11 c G 10 a

a
 46

Uniform-cost search Properties

• What nodes does UCS expand?
• Processes all nodes with cost less than
cheapest solution!
• If that solution costs C* and arcs cost at least
 , then the “effective depth” is roughly C*/
• Takes time O(bC*/) (exponential in effective b c1
…
depth) c2
• How much space does the fringe take?
C*/
“tiers” c3
• Has roughly the last tier, so O(b C*/  )

• Is it complete?
• Assuming best solution has a finite cost and
minimum arc cost is positive, yes!
• Is it optimal?
Yes
Artificial Intelligence Ch3: Problem Solving by Search Prof. Alaa Sagheer
 47

Uniform-cost search Properties

• Remember: UCS explores increasing cost … c1
contours c2
c3

• The good: UCS is complete and optimal!

• The bad:
• Explores options in every “direction” Start Goal
• No information about goal location

[Demo: empty grid UCS

(L2D5)]
• We’ll fix that soon! [Demo: maze with
deep/shallow water
DFS/BFS/UCS (L2D7)]
Artificial Intelligence Ch3: Problem Solving by Search Prof. Alaa Sagheer
 48

Example

Artificial Intelligence Ch3: Problem Solving by Search Prof. Alaa Sagheer

 49

Solution

Artificial Intelligence Ch3: Problem Solving by Search Prof. Alaa Sagheer

 50

Excercise

Artificial Intelligence Ch3: Problem Solving by Search Prof. Alaa Sagheer

 51

Artificial Intelligence Ch3: Solving Problems by Searching Prof. Alaa Sagheer

 52

Depth-first search
❖ The search proceeds immediately to the deepest level of the
search tree, where the nodes have no successors. As those
nodes are expanded, they are dropped from the fringe, so then
the search "backs up" to the next shallowest node that still has
unexplored successors.

Artificial Intelligence Ch3: Problem Solving by Search Prof. Alaa Sagheer

 53

Depth-first search - Example

A
• For example, after searching
A, then B, then D, the search
B C
backtracks and tries another
path from B
D E F G
• Nodes are accessed in the
order A B D E H L M N I O
H I J K PCFGJKQ
• N will be found before J
L M N O P Q • Nodes are tested in the order
DLMNHOPIEBFJQKGCA

Artificial Intelligence Ch3: Problem Solving by Search Prof. Alaa Sagheer

 54

Depth-first search
❑ Implementation:
• This strategy can be implemented by TREE-SEARCH with a
last-in-first-out (LIFO) strategy, also known as a stack.
• As an alternative to the TREE-SEARCH implementation, it is
common to implement depth-first search with a recursive
function that calls itself on each of its children in turn.

Depth-first search has very modest memory

requirements. Once a node has been expanded, it
can be removed from memory as soon as all its
descendants have been fully explored.

Artificial Intelligence Ch3: Problem Solving by Search Prof. Alaa Sagheer

Review: Queue and Stack data structure
- A stack is a limited access data structure
can be viewed as a container of objects that
are inserted and removed according to the
last-in first-out (LIFO) principle.
- Elements can be added and removed from
the stack only at the top. push adds an item
to the top of the stack, pop removes the
item from the top.
- A helpful analogy is to think of a stack of
books; you can remove only the top book,
also you can add a new book on the top.

Artificial Intelligence Ch3: Problem Solving by Search Prof. Alaa Sagheer

 Review: Queue and Stack data structure
- A queue is a container of objects (a linear
collection) that are inserted and removed
according to the first-in first-out (FIFO)
principle.
- Two operations are allowed: enqueue
means to insert an item into the back of
the queue, dequeue means removing the
front item. The picture demonstrates the
FIFO access.
- An excellent example of a queue is a line
of students in the food court. New
additions to a line made to the back of the
queue, while removal (or serving) happens
in the front.

Artificial Intelligence Ch3: Problem Solving by Search Prof. Alaa Sagheer

 57

Depth-first search
• Expand the deepest unexpanded node in the current
fringe.
• Implementation:
• fringe = LIFO queue, i.e., put successors at front
• This strategy can be implemented by TREE-SEARCH with a
last-in-first-out (LIFO) queue, also known as a stack.

Artificial Intelligence Ch3: Problem Solving by Search Prof. Alaa Sagheer

 58

Depth-first search
• Expand the deepest unexpanded node in the current
fringe.
•
• Implementation:
• fringe = LIFO queue, i.e., put successors at front
•

Artificial Intelligence Ch3: Problem Solving by Search Prof. Alaa Sagheer

 59

Depth-first search
• Expand the deepest unexpanded node in the current
fringe.
•
• Implementation:
• fringe = LIFO queue, i.e., put successors at front
•

Artificial Intelligence Ch3: Problem Solving by Search Prof. Alaa Sagheer

 60

Depth-first search
• Expand the deepest unexpanded node in the current
fringe.
•
• Implementation:
• fringe = LIFO queue, i.e., put successors at front
•

Artificial Intelligence Ch3: Problem Solving by Search Prof. Alaa Sagheer

 61

Depth-first search
• Expand the deepest unexpanded node in the current
fringe.
•
• Implementation:
• fringe = LIFO queue, i.e., put successors at front
•

Artificial Intelligence Ch3: Problem Solving by Search Prof. Alaa Sagheer

 62

Depth-first search
• Expand the deepest unexpanded node in the current
fringe.
•
• Implementation:
• fringe = LIFO queue, i.e., put successors at front
•

Artificial Intelligence Ch3: Problem Solving by Search Prof. Alaa Sagheer

 63

Depth-first search
• Expand the deepest unexpanded node in the current
fringe.
•
• Implementation:
• fringe = LIFO queue, i.e., put successors at front
•

Artificial Intelligence Ch3: Problem Solving by Search Prof. Alaa Sagheer

 64

Depth-first search
• Expand the deepest unexpanded node in the current
fringe.
•
• Implementation:
• fringe = LIFO queue, i.e., put successors at front
•

Artificial Intelligence Ch3: Problem Solving by Search Prof. Alaa Sagheer

 65

Depth-first search
• Expand the deepest unexpanded node in the current
fringe.
•
• Implementation:
• fringe = LIFO queue, i.e., put successors at front
•

Artificial Intelligence Ch3: Problem Solving by Search Prof. Alaa Sagheer

 66

Depth-first search
• Expand the deepest unexpanded node in the current
fringe.
•
• Implementation:
• fringe = LIFO queue, i.e., put successors at front
•

Artificial Intelligence Ch3: Problem Solving by Search Prof. Alaa Sagheer

 67

Depth-first search
• Expand the deepest unexpanded node in the current
fringe.
•
• Implementation:
• fringe = LIFO queue, i.e., put successors at front
•

Artificial Intelligence Ch3: Problem Solving by Search Prof. Alaa Sagheer

 68

Depth-first search
• Expand the deepest unexpanded node in the current
fringe.
•
• Implementation:
• fringe = LIFO queue, i.e., put successors at front
•

Artificial Intelligence Ch3: Problem Solving by Search Prof. Alaa Sagheer

 69

Last-in First Out Stack

Artificial Intelligence Ch3: Solving Problems by Searching Prof. Alaa Sagheer

 70

BFS vs. DFS search

• When will BFS outperform DFS?
• When will DFS outperform BFS?

Artificial Intelligence Ch3: Solving Problems by Searching Prof. Alaa Sagheer

 71

Depth-First vs. Breadth-First

▪ When a breadth-first search succeeds, it finds a minimum-
depth (nearest the root) goal node
▪ When a depth-first search succeeds, the found goal node is not
necessarily minimum depth

▪ For a large tree, breadth-first search memory requirements may

be excessive
▪ For a large tree, a depth-first search may take an excessively
long time to find even a very nearby goal node

How can we combine the advantages and avoid the

disadvantages of these two search techniques?

Artificial Intelligence Ch3: Problem Solving by Search Prof. Alaa Sagheer

 72

Depth-limited search
• The problem of unbounded trees can be alleviated by supplying
depth-first search with a predetermined depth limit l . That is,
nodes at depth l are treated as if they have no successors.
• Sometimes, depth limits can be based on knowledge of the
problem.
For example, on the map of Romania there are 20 cities.
Therefore, we know that if there is a solution, it must be of
length 19 at the longest, so l = 19 is a possible choice.
But in fact if we studied the map carefully, we would discover
that any city can be reached from any other city at most 10
steps (the diameter of the state space), which leads to a more
efficient depth-limited search.
• For most problems, however, we will not know a good depth
limit until we have solved the problem.
Artificial Intelligence Ch3: Problem Solving by Search Prof. Alaa Sagheer
 73

Depth-limited search

Artificial Intelligence Ch3: Problem Solving by Search Prof. Alaa Sagheer

 74

BFS vs. DFS search

• When will BFS outperform DFS?
• When will DFS outperform BFS?
• Comparing BFS and DFS, the big advantage of DFS is that it has much lower
memory requirements than BFS, because it’s not necessary to store all of the
child pointers at each level.
• Depending on the data and what you are looking for, either DFS or BFS could be
advantageous.
For example
Given a family tree, if one were looking for someone on the tree who’s still alive, then it would be
safe to assume that person would be on the bottom of the tree. This means that a BFS would
take a very long time to reach that last level. A DFS, however, would find the goal faster.
But, if one were looking for a family member who died a very long time ago, then that person
would be closer to the top of the tree.

Artificial Intelligence Ch3: Solving Problems by Searching Prof. Alaa Sagheer

 75

Iterative deepening search

• The iterative deepening search algorithm, which repeatedly applies depth-
limited search with increasing limits.
• It does this by increasing the limit gradually, first 0, second 1, third 2, and so on
until a goal is found.
• This (i.e. goal is found) will occur when the depth limit reaches d, the depth of
the shallowest goal node.
• Algorithm is below, it combines the benefits of depth-first and breadth-first
search,
• Like depth-first search, its memory requirements are very modest,
• Like breadth-first search, it is complete when the branching factor is finite and
optimal as well.
• It terminates when a solution is found or if the depth- limited search returns
failure, meaning that no solution exists.

Artificial Intelligence Ch3: Problem Solving by Search Prof. Alaa Sagheer

 76

Iterative deepening search / =0

Artificial Intelligence Ch3: Problem Solving by Search Prof. Alaa Sagheer

 77

Iterative deepening search / =0

Artificial Intelligence Ch3: Problem Solving by Search Prof. Alaa Sagheer

 78

Iterative deepening search / =1

Artificial Intelligence Ch3: Problem Solving by Search Prof. Alaa Sagheer

 79

Iterative deepening search / =1

Artificial Intelligence Ch3: Problem Solving by Search Prof. Alaa Sagheer

 80

Iterative deepening search / =1

Artificial Intelligence Ch3: Problem Solving by Search Prof. Alaa Sagheer

 81

Iterative deepening search / =1

Artificial Intelligence Ch3: Problem Solving by Search Prof. Alaa Sagheer

 82

Iterative deepening search / =1

Artificial Intelligence Ch3: Problem Solving by Search Prof. Alaa Sagheer

 83

Iterative deepening search / =2

Artificial Intelligence Ch3: Problem Solving by Search Prof. Alaa Sagheer

 84

Iterative deepening search / =2

Artificial Intelligence Ch3: Problem Solving by Search Prof. Alaa Sagheer

 85

Iterative deepening search / =2

Artificial Intelligence Ch3: Problem Solving by Search Prof. Alaa Sagheer

 86

Iterative deepening search / =2

Artificial Intelligence Ch3: Problem Solving by Search Prof. Alaa Sagheer

 87

Iterative deepening search / =2

Artificial Intelligence Ch3: Problem Solving by Search Prof. Alaa Sagheer

 88

Iterative deepening search / =2

Artificial Intelligence Ch3: Problem Solving by Search Prof. Alaa Sagheer

 89

Iterative deepening search / =2

Artificial Intelligence Ch3: Problem Solving by Search Prof. Alaa Sagheer

 90

Iterative deepening search / =2

Artificial Intelligence Ch3: Problem Solving by Search Prof. Alaa Sagheer

 91

Iterative deepening search / =2

Artificial Intelligence Ch3: Problem Solving by Search Prof. Alaa Sagheer

 92

Iterative deepening search l =1

Artificial Intelligence Ch3: Problem Solving by Search Prof. Alaa Sagheer

 93

Iterative deepening search l =2

Artificial Intelligence Ch3: Problem Solving by Search Prof. Alaa Sagheer

 94

Iterative deepening search l =3

Solution

• It may seem wasteful, because states are generated multiple times.

• However, this is not very costly! The reason is that in a search tree with the same branching
factor at each level, most of the nodes are in the bottom level, so it does not matter much that
the upper levels are generated multiple times.

Artificial Intelligence Ch3: Problem Solving by Search Prof. Alaa Sagheer

 95

Iterative deepening search

• Idea: get DFS’s space advantage with BFS’s time /
shallow-solution advantages
• Run a DFS with depth limit 1. If no solution… b
…
• Run a DFS with depth limit 2. If no solution…
• Run a DFS with depth limit 3. …..

• Isn’t that wastefully redundant?

• Generally most work happens in the lowest level
searched, so not so bad!

Artificial Intelligence Ch3: Problem Solving by Search Prof. Alaa Sagheer

 96

Iterative deepening vs. Breadth first

In an iterative deepening search, the nodes on the bottom level (depth d) are
generated once, those on the next to bottom level are generated twice, and so on, up
to the children of the root, which are generated d times. So the total number of nodes
generated is:

N(IDS) = (d) b + (d- l) b2 +...+(1)bd,

N(BFS) = b + b2 + b3 +….+ bd + (bd+1- b)
• The former shows time complexity of O(bd) where the second is O(bd+1),
• Iterative deepening is FASTER than Breadth First search, despite the repeated
generation of states. For example, if b =10 and d =5, then we have:
N (IDS) = 50+400+3.000+20.000+100.000 = 123.450
N(BFS) = 10+100+1.000+10.000+100.000+999.990=1,111,100

Iterative deepening is the preferred uninformed search method when

there is a large search space and the depth of the solution is not known.

Artificial Intelligence Ch3: Problem Solving by Search Prof. Alaa Sagheer

 97

Iterative deepening vs. Depth-first

❖Time requirements on binary tree with 7 levels
Nodes at Nodes searched Nodes searched
each level by DFS by iterative IDS
1 1 1
2 +2 = 3 +3 = 4
4 +4 = 7 +7 = 11
8 +8 = 15 +15 = 26
16 +16 = 31 +31 = 57
32 +32 = 63 +63 = 120
64 +64 = 127 +127 = 247
128 +128 = 255 +255 = 502
Artificial Intelligence Ch3: Problem Solving by Search Prof. Alaa Sagheer
 98

Iterative deepening vs. Depth-first

❖ Time requirements for depth-first iterative deepening on branching factor 4

Nodes at Nodes searched Nodes searched

each level by DFS by iterative IDS
1 1 1
4 +4 = 5 +5 = 6
16 +16 = 21 +21 = 27
64 +64 = 85 +85 = 112
256 +256 = 341 +341 = 453
1024 +1024 = 1365 +1365 = 1818
4096 +4096 = 5461 +5461 = 7279
16384 +16384 = 21845 +21845 = 29124

Artificial Intelligence Ch3: Problem Solving by Search Prof. Alaa Sagheer

 99

Iterative deepening vs. Depth-first

• When searching a binary tree to depth 7:
• DFS requires searching 255 nodes
• IDS requires searching 502 nodes
• IDS takes only about twice as long

• When searching a tree with branching factor of 4 (each node

may have four children):
• DFS requires searching 21845 nodes
• IDS requires searching 29124 nodes
• IDS takes about 4/3 = 1.33 times as long

The higher the branching factor, the lower the relative cost
of iterative deepening depth first search
Artificial Intelligence Ch3: Problem Solving by Search Prof. Alaa Sagheer
 100

Last-in First Out with depth limit

Artificial Intelligence Ch3: Solving Problems by Searching Prof. Alaa Sagheer

 101

Avoiding Repeated States

• There is a possibility of wasting time by expanding states that have already
been encountered and expanded before.
• Failure to detect repeated states can turn a linear problem into an exponential

• If an algorithm remembers every state that it has visited, then it can be viewed
as exploring the state-space graph directly.

• We can modify the general TREE-SEARCH algorithm to include

a data structure called the closed list, which stores every
expanded node.
• If the current node matches a node on the closed list, it is discarded instead of

being expanded.

•Artificial Intelligence Ch3: Problem Solving by Search Prof. Alaa Sagheer

 102

SUMMARY
• This chapter has introduced methods that an agent can use to select actions in environments
that are deterministic, observable, static, and completely known. In such cases, the: agent can
construct sequences of actions that achieve its goals; this process is called search.
• Before an agent can start searching for solutions, it must formulate a goal and then use the
goal to formulate a problem.
• A problem consists of four parts: the initial state, a set of actions, a goal test function, and
a path cost function. The environment of the problem is represented by a state space. A
path through the state space from the initial state to a goal state is a solution.
• Search algorithms are judged on the basis of completeness, optimality, time complexity,
and space complexity. Complexity depends on b, the branching factor in the state space,
and d, the depth of the shallowest solution.
• Breadth-first search selects the shallowest unexpanded node in the search tree for
expansion. It is complete, optimal for unit step costs, and has time and space complexity of
O(bd+1). The space complexity makes it impractical in most cases.

Artificial Intelligence Ch3: Solving Problems by Searching Prof. Alaa Sagheer

 103

SUMMARY
• Uniform-cost search is similar to breadth-first search but expands the node with lowest
path cost, g (n) . It is complete and optimal if the cost of each step exceeds some positive
bound.
• Depth-first search selects the deepest unexpanded node in the search tree for expansion. It
is neither complete nor optimal, and has time complexity of O(bm) and space complexity of
O(bm ), where m is the maximum depth of any path in the state space.
• Depth-limited search imposes a fixed depth limit on a depth-first search.
• Iterative deepening search calls depth-limited search with increasing limits until a goal is
found. It is complete, optimal for unit step costs, and has time complexity of O(bd) and
space complexity of O(bd).

Artificial Intelligence Ch3: Solving Problems by Searching Prof. Alaa Sagheer

 104

Exercises
Q1: What are the properties of a search problem?

Fully observable, Deterministic,

Sequential, Static, Discrete, and Single
Artificial Intelligence Ch3: Solving Problems by Searching Prof. Alaa Sagheer
 105

Exercises
Q1: What is the solution of the Vacuum problem?

Artificial Intelligence Ch3: Solving Problems by Searching Prof. Alaa Sagheer

 Exercises
Please try to solve the following exercises:
3.1
3.2
3.6
3.7-a
3.8- a,b
3.10-3.13

Assoc.Prof. Alaa Sagheer ‫ عالء الصغيـر‬.‫د‬

106
[email protected]
Artificial Intelligence Ch2: Intelligent Agents Prof. Alaa Sagheer
Assoc.Prof. Alaa Sagheer ‫ عالء الصغيـر‬.‫د‬
[email protected]
107