LectureWeek2 93479

Artificial Intelligence
CU6051NI
Lecture 2 - Search Agents

BSc (Hons) Computing
Week 2
Instructor: Sarun Dahal

Sarun.dahal@islingtoncollege.edu.np
Learning objectives
➢ Familiar with different agent-programs
➢ Get acquainted with various types of searching problems
➢ Describe strategies (framework) for formulating searching problems
➢ Be able to describe uninformed search algorithms (BFS and DFS)
CU6051NI Artificial Intelligence January 17, 2024 2

Agent programs
Goal-based agents
 Knowing something about the current state of environment is not always sufficient
o Example, what action a taxi should take (turn left, right, go straight) at junction?
 With current state description, agents need a goal information describing desirability.
 Agents consider the impact of actions on future states and work towards a goal.

Agent programs
Goal-based agents
 Agents tries to get answers of the following questions during decision making
1. What will happen if I do such-and-such? And
2. Will that make me happy? (help to achieve my goals)
 Searching and planning methods help to find sequence of actions to achieve goals
 We will be learning about only searching methods in the module.

Example of Searching Problems
Examples

Examples

Examples

Examples
Explore
Examples
The 8-queen problem: on a chess board, place 8 queens so that no queen is attacking any
other horizontally, vertically or diagonally.

Formulation of Search Problems
Problem Solving Agent
 Intelligent agents try to maximize their performance measure
 One way to maximize the performance measure is if an agent can adopt a goal and
aim at satisfying it.
 How do you define a problem?

o Goal formulation
o Problem formulation

Formulation of Search Problem
 Goal formulation
o Goals help to organize agent’s behavior to achieve its goal.
o Agent’s task is to find possible actions (now & in future) to achieve the goal.
o Is first step and considers the current situation and performance measure.
o A goal is a set of world state where the goal is satisfied.
Example: if the task is to travel from one town to another to reach a destination,
then arriving at each town along the way is an example of a state.
 Problem formulation
o Is a process to decide what actions and state to consider, given a goal.

 Agent: entity that perceives its environment and acts upon that environment.
 State: a configuration of an agent and its environment

 A search problem can be formulated in five steps
Initial state - the state in which an agent starts

Actions – choices that can be made in a state
o ACTIONS(s) returns set of actions that can be executed in state s

Transition model – a description of what state results from performing any applicable
action in any state
o RESULT(s, a) returns the state resulting from performing action a in state s

State space - the set of all states reachable from the initial state by any sequence of
actions

Goal test – A way to determine whether a given state is a goal state.
Path cost – Numerical cost associated with a given path
Agent chooses a cost function that reflects its own performance measure.

 Search Problem
o Initial state
o Action
o Transition model
o Goal test
o Path cost function
 Solution
o A sequence of actions that lead from the initial state to a goal state.
 Optimal solution
o A solution that gives lowest path cost among all solutions.

Examples
 States: all arrangements of 0 to 8 queens on the board.
 Initial state: No queen on the board
 Actions: Add a queen to any empty square
 Transition model: updated board
 Goal test: 8 queens on the board with none attacked

Examples
start state goal state

Examples
 States: Location of each of the 8 tiles in
the 3x3 grid
 Initial state: Any state
 Actions: Move Left, Right, Up or Down
 Transition model: Given a state and an
action, returns resulting state
 Goal test: state matches the goal state?
 Path cost: total moves, each move costs 1.

Examples
Find the shortest route

between 2 cities

Examples
 States: {Los Angeles, New York,
Denver,…}
 Initial state: In Boston
 Actions: Go New York, etc.
 Transition model:
Results(In(Boston),Go(New
York)) = In(New York)
 Goal test: In(Denver)
 Path cost: path length in kms

Real-world examples
• Route finding problem: going from one location to another, think google maps

Real-world examples
• Travelling salesperson problem: Find the shortest tour to visit each city exactly once

State space vs. search space
 State space: a physical configuration
 Search space: an abstract configuration represented by a search tree or graph of
possible solutions.
 Search tree: models the sequence of actions
o Root: initial state
o Branches: actions
o Nodes: results from actions. A node has: parent, children, depth, path cost,
associated state in the state space.
 Expand: A function that given a node, creates all children nodes

Search Space Regions
 The search space is divided into three regions:
1. Explored (a.k.a. Closed List, Visited Set)
2. Frontier (a.k.a. Open List, the Fringe)
3. Unexplored.
 The essence of search is moving nodes from regions (3) to (2) to (1), and the essence
of search strategy is deciding the order of such moves.
 In the following we adopt the following color coding:
o orange nodes are explored,
o grey nodes are the frontier,
o white nodes are unexplored, and
o black nodes are failures.
Searching Algorithms
Examples of search agents
Let's show the first steps in growing the search tree to find a route from San Francisco to
another city


o white nodes are unexplored,




Search Strategies
 A strategy is defined by picking the order of node expansion
 Strategies are evaluated along the following dimensions:
o Completeness - Does it always find a solution if one exists?
o Time complexity - Number of nodes generated/expanded
o Space complexity - Maximum number of nodes in memory
o Optimality - Does it always find a least-cost solution?
 Searching strategies
o Uninformed search – Given no information about the problem but definition
o Informed search – Given some knowledge or guidance where to look for solution

Uninformed search
 Use no domain knowledge!
 Strategies
o Breadth-first search (BFS): Expand shallowest node
o Depth-first search (DFS): Expand deepest node

Breadth-first search (BFS)
BFS: Expand shallowest node first

Frontier: C, D, E
Frontier: A (root ) Frontier: B, C

Frontier: D, E, F, G Frontier: E, F, G, H, I
Frontier: C, D, E

Frontier: E, F, G, H, I Frontier: F, G, H, I, J, K Frontier: G, H, I, J, K, L, M

implementation:
FIFO (Queue)
Frontier: G, H, I, J, K, L, M Frontier: H, I, J, K, L, M, N, O Frontier: I, J, K, L, M, N, O

Depth-first Search (DFS)
DFS: Expand deepest node first

Frontier: A (root) Frontier: B, C Frontier: D, E, C

Frontier: D, E, C Frontier: H, I, E, C Frontier: I, E, C

Frontier: I, E, C Frontier: E, C Frontier: J, K, C

implementation:
DFS: Expand deepest node first LIFO (Stack)
Frontier: J, K, C Frontier: K, C Frontier: C


Exercises
Exercise
Question: What is the order of visits of the nodes and the path returned by BFS and DFS?
S – start
G – goal
ignore the numbers in
the path for now

Exercise: BFS
Exercise: BFS

Exercise: BFS

Exercise: BFS

Exercise: BFS

Exercise: BFS

Exercise: BFS

Exercise: BFS

Exercise: BFS

Exercise: BFS

Exercise: DFS
Exercise: DFS

Exercise: DFS

Exercise: DFS

Exercise: DFS

Exercise: DFS

Exercise: DFS

Exercise: DFS

Exercise: DFS

Exercise: DFS

Summary
Summary
• Discussed the structures of agent.
• Acquainted with different searching problems.
• Learned how to formulate search problems systematically.
• Understood the DFS and BFS search algorithms, uninformed searching strategies.

End of Lecture 2

Thank you !
Any questions ?

LectureWeek2 93479

Uploaded by

Document Information

Original Description:

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

LectureWeek2 93479

Uploaded by

Copyright:

Available Formats

Artificial Intelligence

Lecture 2 - Search Agents

Instructor: Sarun Dahal

CU6051NI Artificial Intelligence January 17, 2024 2

CU6051NI Artificial Intelligence January 17, 2024 3

 We will be learning about only searching methods in the module.

CU6051NI Artificial Intelligence January 17, 2024 4

CU6051NI Artificial Intelligence January 17, 2024 6

CU6051NI Artificial Intelligence January 17, 2024 7

CU6051NI Artificial Intelligence January 17, 2024 8

CU6051NI Artificial Intelligence January 17, 2024 10

 How do you define a problem?

CU6051NI Artificial Intelligence January 17, 2024 12

CU6051NI Artificial Intelligence January 17, 2024 13

CU6051NI Artificial Intelligence January 17, 2024 14

CU6051NI Artificial Intelligence January 17, 2024 15

CU6051NI Artificial Intelligence January 17, 2024 16

CU6051NI Artificial Intelligence January 17, 2024 17

CU6051NI Artificial Intelligence January 17, 2024 18

CU6051NI Artificial Intelligence January 17, 2024 19

CU6051NI Artificial Intelligence January 17, 2024 20

CU6051NI Artificial Intelligence January 17, 2024 21

start state goal state

CU6051NI Artificial Intelligence January 17, 2024 22

CU6051NI Artificial Intelligence January 17, 2024 23

Find the shortest route

CU6051NI Artificial Intelligence January 17, 2024 24

CU6051NI Artificial Intelligence January 17, 2024 25

CU6051NI Artificial Intelligence January 17, 2024 26

CU6051NI Artificial Intelligence January 17, 2024 27

CU6051NI Artificial Intelligence January 17, 2024 28

CU6051NI Artificial Intelligence January 17, 2024 31

o orange nodes are explored,

o black nodes are failures.

CU6051NI Artificial Intelligence January 17, 2024 32

o orange nodes are explored,

o black nodes are failures.

CU6051NI Artificial Intelligence January 17, 2024 33

o black nodes are failures.

CU6051NI Artificial Intelligence January 17, 2024 34

CU6051NI Artificial Intelligence January 17, 2024 35

CU6051NI Artificial Intelligence January 17, 2024 36

CU6051NI Artificial Intelligence January 17, 2024 38

CU6051NI Artificial Intelligence January 17, 2024 39

CU6051NI Artificial Intelligence January 17, 2024 40

Frontier: E, F, G, H, I Frontier: F, G, H, I, J, K Frontier: G, H, I, J, K, L, M

CU6051NI Artificial Intelligence January 17, 2024 41

Frontier: G, H, I, J, K, L, M Frontier: H, I, J, K, L, M, N, O Frontier: I, J, K, L, M, N, O

CU6051NI Artificial Intelligence January 17, 2024 42

CU6051NI Artificial Intelligence January 17, 2024 44

Frontier: A (root) Frontier: B, C Frontier: D, E, C

CU6051NI Artificial Intelligence January 17, 2024 45

Frontier: D, E, C Frontier: H, I, E, C Frontier: I, E, C

CU6051NI Artificial Intelligence January 17, 2024 46

Frontier: I, E, C Frontier: E, C Frontier: J, K, C

CU6051NI Artificial Intelligence January 17, 2024 47

Frontier: J, K, C Frontier: K, C Frontier: C

CU6051NI Artificial Intelligence January 17, 2024 48

CU6051NI Artificial Intelligence January 17, 2024 49

CU6051NI Artificial Intelligence January 17, 2024 51

CU6051NI Artificial Intelligence January 17, 2024 53

CU6051NI Artificial Intelligence January 17, 2024 54