TAI2151 – ARTIFICIAL INTELLIGENCE

FUNDAMENTALS

LECTURE 5 – SOLVING
PROBLEMS BY SEARCHING
(INFORMED SEARCHES)
PART 1

Solving Problems By Searching – Informed Searches

 LECTURE OUTLINE

• Part 1: Best First Search

• Greedy search
• A* search
• Part 2: Heuristic functions

Solving Problems By Searching – Informed Searches

 INFORMED SEARCH

• Problems:
• How to achieve goal?
• Goal: get a solution, get the best solution
• How to maximise the utility/profit?
• Ideally, we want AI to know the knowledge and draw
conclusion
• Knowledge: a sequence of action that I took
• Conclusion: predict the result from the actions

Solving Problems By Searching – Informed Searches

 INFORMED SEARCH
• Search strategy that uses problem
specific knowledge beyond the
definition of the problem itself
• To fine a goal with less search
space (that is informed)
• Can find solutions more efficiently

Solving Problems By Searching – Informed Searches

 INFORMED VS UNINFORMED SEARCH

• Uninformed search:
• Use only available information in the problem
• For e.g.:
• From City A to D: look only the available path from A to D

• Informed search:
• Use knowledge specific to a particular problem
• Interpret the problem and available information
• Involve reasoning
• For e.g.:
• From City A to D: getting the shortest path from A to D, is better than the longest path from A
to D; or find the path with least traffic jam

Solving Problems By Searching – Informed Searches

 PROBLEM: FINDING A MINIMUM COST PATH
• Previously we wanted an arbitrary path to a goal.
• Now, we want the minimum cost path to a goal G
• Cost of a path = sum of individual transitions along path
• How to obtain the path cost? By defining a heuristic function, commonly write as h(n)
• Heuristics: serving to aid discovery
• An estimation function that estimates how good a node is
• “Good”, define based on domain specific information. For e.g.: could be the shortest path in a
map, number of pieces moved for 8-puzzle
• Examples of path-cost:
• Navigation
• path-cost g(n) = distance to node (n) from start node in miles
• minimum => minimum time, least fuel, least distance

• 8-Puzzle
• path-cost = number of pieces moved
• minimum => least time to solve the puzzle

Solving Problems By Searching – Informed Searches

 TERMINOLOGY
• Agent: AI or someone who try to solve the problem
• States: configuration of the agent in its environment.
• Initial state: initial starting state
• Goal state: solution state
• Operators/Actions: a kind of (movement) that make within a state
(choices that can be made within a state)
• Goal test: way to determine whether a given state is a goal state
• A machine needs some way to encode whether a state happens to be in
is the goal state
• Path Cost: numerical cost associated with a given path
• Transition Model: 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

Solving Problems By Searching – Informed Searches

 INFORMED SEARCHES

 Best-First Search
* Greedy Search
* A* Search

 Heuristic Functions
* Designing a Heuristic Function

Solving Problems By Searching – Informed Searches

 BEST FIRST SEARCH

• A search algorithm that expands the node (from one state

to another state) that is closest to the goal
• We use a heuristic function to make estimation (we guess,
but guess with mathematics, i.e. probability and statistics
logic rules & facts)

Solving Problems By Searching – Informed Searches

 Informed search strategy uses problem-specific knowledge

Solving Problems By Searching – Informed Searches

 GREEDY SEARCHES

 Search algorithm: expands node that is closest

to the goal

 Estimate through heuristic function (a function

that calculates an approximate cost (ranks
alternatives) in search algorithm at each
branching step based on available information
to decide which branch to follow)

 Since it is approximate, so not necessary it will

be the best solution (most optimize)

Solving Problems By Searching – Informed Searches

 Minimize estimated cost to reach a goal: Greedy search

Solving Problems By Searching – Informed Searches

 GREEDY SEARCH
• Let evaluation function h(n) be an estimate of cost from node n
to goal; this function is often called a heuristic function
(heuristic means to find or to discover) and is denoted by h(n).
e.g., hSLD(n) = straight-line distance from n to Bucharest
• Greedy search expands the node that appears to be closest to
goal
• Contrast with uniform-cost search in which lowest cost path
from start is expanded
• Greedy search resembles depth-first search since it prefers to
follows a single path all the way to the goal, but it will backup
when it hits a dead end. So
• it is not complete
• it is not optimal
• time complexity is O(bm) and space complexity is O(bm)

Solving Problems By Searching – Informed Searches

Map of Romania with road distances in km, and straight-line distances to Bucharest.
Goal: Bucharest
Initial state: Arad
States: various cities
Actions: drive between cities or choose
Minimizenext
estimated
city cost to reach a goal: Greedy Search
Solution: Sequence of cities to reach
Bucharest

Solving Problems By Searching – Informed Searches

Solving Problems By Searching – Informed Searches
 GREEDY SEARCH - EXAMPLE

Solving Problems By Searching – Informed Searches

 GREEDY SEARCH – EXAMPLE (CONT.)

Solving Problems By Searching – Informed Searches

 GREEDY SEARCH – EXAMPLE (CONT.)

Arad → Sibiu → Fagaras → Bucharest : not optimal (140 + 99 +211 = 450)

( 32 kilometer longer than Arad → Sibiu → Rimnicu → Pitesti → Bucharest): Optimal
(140 + 80 + 97 + 101 = 418)
Solving Problems By Searching – Informed Searches
 PROPERTIES OF GREEDY SEARCH
• Complete:
No:
• Fails in infinite-depth spaces -- example
• Spaces with loops (e.g. going from Lasi to Oradea)
→ Complete in finite spaces with repeated state
checking
• Optimal:
No - optimal path goes through Pitesti.
O(bm): like depth first, but a good heuristic function
• Time:
gives dramatic improvement on average
O(bm): Potentially keeps all nodes in memory
• Space:

Let b: Branching factor

m: Maximum Depth
Solving Problems By Searching – Informed Searches
 A*- a special Best-first search
• Goal: find a minimum total cost path
• Notation:
– g(n) = cost of current path from start to node n in the search tree.
– h(n) = estimate of the cheapest cost of a path from n to a goal.
– evaluation function: f = g + h
• f(n) estimates the cheapest cost solution path that goes through n.
– h*(n) is the true cheapest cost from n to a goal.
– g(n) is the true shortest path from the start s, to n.

• If the heuristic function, h(n), always underestimate the true cost (h(n)
is smaller than h*(n)), then A* is guaranteed to find an optimal
solution.

Solving Problems By Searching – Informed Searches

 MINIMIZING THE TOTAL PATH COST: A* IDEA

Solving Problems By Searching – Informed Searches

 A* Search

Straight-line distance is admissible because the shortest

path between any two points is a straight line.

Solving Problems By Searching – Informed Searches

 EXAMPLE – CONT.

Solving Problems By Searching – Informed Searches

 EXAMPLE – CONT.

Solving Problems By Searching – Informed Searches

 EXAMPLE – FINALLY…

Solving Problems By Searching – Informed Searches

 PROPERTIES OF A*
• Complete: Yes, unless there are infinitely many
nodes with f  f(G)
• Optimal: Yes
• Time: Exponential with path length
• Space: Potentially keeps all nodes in memory
(runs out of space before it runs out of time)

Note, A* is also optimally efficient

for a given heuristic: That is, no other
optimal search algorithm is guaranteed to
open fewer nodes. Let b: Branching factor
m: Maximum Depth
Solving Problems By Searching – Informed Searches