0% found this document useful (0 votes)

131 views

Week 05 Lecture 01 AStar Reading Assignment

The document discusses different types of search algorithms used in artificial intelligence. It introduces informed search which uses domain knowledge in the form of a heuristic function to estimate how close a state is to the goal. It provides examples of informed search strategies like greedy best-first search, A* search, and IDA* search which make use of heuristics. The document also discusses properties of admissible heuristics and how A* search is optimal if the heuristic is admissible.

Uploaded by

Osama Adeel

We take content rights seriously. If you suspect this is your content, claim it here.

Available Formats

Download as PDF, TXT or read online on Scribd

0% found this document useful (0 votes)

131 views

Week 05 Lecture 01 AStar Reading Assignment

Uploaded by

Osama Adeel

We take content rights seriously. If you suspect this is your content, claim it here.

Available Formats

Download as PDF, TXT or read online on Scribd

You are on page 1/ 87

Artificial Intelligence

Search Agents
Informed search
Informed search

Use domain knowledge!

• Are we getting close to the goal?

• Use a heuristic function that estimates how close a state is to

the goal

• A heuristic does NOT have to be perfect!

Informed search

Use domain knowledge!

• Are we getting close to the goal?

• Use a heuristic function that estimates how close a state is to

the goal

• A heuristic does NOT have to be perfect!

• Example of strategies:

1. Greedy best-first search

2. A* search

3. IDA*
Informed search

Heuristic!

The distance is the straight line distance. The goal is to get to Sault Ste Marie,
so all the distances are from each city to Sault Ste Marie.
Greedy search

• Evaluation function h(n) (heuristic)

• h(n) estimates the cost from n to the closest goal

• Example: hSLD(n) = straight-line distance from n to Sault Ste

Marie

• Greedy search expands the node that appears to be closest to

goal
Greedy search
Greedy search example
The initial state:
Greedy search example
After expanding St Louis:
Greedy search example
After expanding Chicago:
Greedy search example
After expanding Duluth:
Examples using the map
Start: Saint Louis
Goal: Sault Ste Marie

Greedy search
Examples using the map
Start: Las Vegas
Goal: Calgary

Greedy search
A* search

• Minimize the total estimated solution cost

• Combines:
– g(n): cost to reach node n
– h(n): cost to get from n to the goal
– f (n) = g(n) + h(n)

f (n) is the estimated cost of the cheapest

solution through n
A* search
A* search example
The initial state:
A* search example
After expanding St Louis:
A* search example
After expanding Chicago:
A* search example
After expanding Kansas City:
A* search example
After expanding Little Rock:
A* search example
After expanding Nashville:
A* search example
After expanding Pittsburgh:
A* search example
After expanding Toronto:
Examples using the map
Start: Saint Louis
Goal: Sault Ste Marie

A*
Examples using the map
Start: Las Vegas
Goal: Calgary

A*
Admissible heuristics

A good heuristic can be powerful.

Only if it is of a “good quality”

Admissible heuristics

A good heuristic can be powerful.

Only if it is of a “good quality”

A good heuristic must be admissible.

Admissible heuristics

• An admissible heuristic never overestimates the cost to reach

the goal, that is it is optimistic

• A heuristic h is admissible if

8 node n, h(n)  h⇤(n)

where h⇤ is true cost to reach the goal from n.

• hSLD (used as a heuristic in the map example) is admissible

because it is by definition the shortest distance between two
points.
A* Optimality
If h(n) is admissible, A* using tree search is optimal.
A* Optimality
If h(n) is admissible, A* using tree search is optimal.
Rationale:
• Suppose Go is the optimal goal.
• Suppose Gs is some suboptimal goal.
• Suppose n is an unexpanded node in the fringe such that n is
on the shortest path to Go.
• f (Gs) = g(Gs) since h(Gs) = 0
f (Go) = g(Go) since h(Go) = 0
f (Gs) > g(Go) since Gs is suboptimal
Then f (Gs) > f (Go) . . . (1)
• h(n)  h⇤(n) since h is admissible
g(n) + h(n)  g(n) + h⇤(n) = g(Go) = f (Go)
Then, f (n)  f (Go) . . . (2)
From (1) and (2) f (Gs) > f (n)
so A* will never select Gs during the search and hence A* is
optimal.
A* search criteria

• Complete: Yes

• Time: exponential

• Space: keeps every node in memory, the biggest problem

• Optimal: Yes!
Heuristics

• The solution is 26 steps long.

• h1(n) = number of misplaced tiles
• h2(n) =total Manhattan distance (sum of the horizontal and
vertical distances).
• h1(n) = 8
• Tiles 1 to 8 in the start state gives: h2 = 3 + 1 + 2 + 2 + 2 +
3 + 3 + 2 = 18 which does not overestimate the true solution.
Search Algorithms: Recap

• Uninformed Search: Use no domain knowledge.

BFS, DFS, DLS, IDS, UCS.

• Informed Search: Use a heuristic function that estimates how

close a state is to the goal.