Advanced
Knowledge
Representation
Section 4
Uniform Cost Search
• It is a modification for BFS algorithm.
• Use a priority queue (least cost first)
• Pop the element with least cost
• If two elements have the same cost, use the alphabetic order
Uniform Cost Search (Example
1)
A 4
2 G
2
C
S 5
1
5
3 D
B 4
Uniform Cost Search
• Solution
Priority Queue Current
]S[
]0[
]S,D[ ]S,B[ ]S,A[ ]S[
]5[ ] 3[ ] 2 [ ]0[
]S,A,C[ ]S,D[ ]S,B[ ]S,A[
]6[ ] 5[ ] 3 [ ]2[
]S,B,D[ ]S,A,C[ ]S,D[ ]S,B[
]7[ ] 6[ ] 5[ ]3[
]S,D,G[ ]S,B,D[ ]S,A,C[ ]S,D[
]10[ ] 7[ ] 6[ ]5[
]S,A,C,D[ ]S,D,G[ ]S,B,D [ ]S,A,C[
]S,A,C,G[ ]6[
]8[ ] 7[ ]10[ ] 7[
Uniform Cost Search (Example
2)
Uniform Cost Search (Example
2)
• Solution
Priority Queue Current
]a[
]0[
]a,c[ ]a,b[ ]a[
]3[ ]6[ ]0[
]a,c,e[ ]a,c,d[ ]a,b[ ]a,c[
]5[ ]11[ ]6[ ]3[
]a,c,e,d[ ]a,c,d[ ]a,b[ ]a,c,e[
]14[ ]11[ ]6[ ]5[
]a,b,d[ ]a,c,e,d[ ]a,c,d[ ]a,b[
] 8[ ]14[ ]11[ ]6[
]a,b,d[
]8[
A Star Search
• Solution Idea
• Use a priority queue (like Uniform-cost-search).
• A* algorithm extends the path that minimizes the following function:
f(n)=g(n)+h(n)
• where n is the last node in the path, g(n) is the cost of the path from start node to
node n, h(n) is a heuristic function that estimates cost of the cheapest path
from node n to the goal node.
A Star Search
• Pop the element with least F-cost.
• If two elements have the same F-cost, use alphabetic order.
• Remember: If a node has been visited, we don’t expand it again.
A Star Search (Example 1)
• Use algorithm to find the path between S and G
Heuristic Data
H Node
A 12
1 S
7
5
S 2 A
C G 6
3
4 B
4 2
B
2 C
0 G
A Star Search (Example 1)
• Definitions
Assume the following path S->A->C:
• G-Cost: The cost to move from S to A to C (G-cost=1+5=6)
• H-Cost: The H value for the last node on the path C from the given table (H-
Cost=2)
• F-Cost: G-Cost+ H-Cost=6+2=8 ()
A Star Search (Example 1)
• Solution
Priority Queue Current
The Solution is
]s[
]7[ S->A->B->C->G
]S,B[ ]S,A[ ]s[
]8[ ] 7[ ]7[
]S,A,G[ ]S,A,C[ ]S,A,B[ ]S,B[ ]S,A[
]13[ ] 8[ ] 7[ ] 8[ ]7 [
]S,A,B,C[ ]S,A,G[ ]S,A,C[ ]S,B[ ]S,A,B[
] 7[ ]13[ ] 8[ ] 8[ ]7 [
]S,A,B,C,G[ ]S,A,G[ ]S,A,C[ ]S,B[ ]S,A,B,C[
]8[ ]13[ ] 8[ ] 8[ ]7[
]S,A,B,C,G[
]8[
A Star Search (Example 2)
• Find the most cost-effective path to reach from start node A to the
final node J.
A Star Search (Example 2)
• The numbers written on edges represent the distance between the
nodes.
• The numbers written on nodes represent the heuristic value.
A Star Search (Example 2)
• Solution
Priority Queue Current
The Solution is
]A[
]10[ A->F->G->I->J
]A,F[ ]A,B[ ]A[
]9[ ]14[ ]10[
]A,F,H[ ]A,F,G[ ]A,B[ ]A,F[
]13[ ]9[ ]14[ ]9 [
]A,F,G,I[ ]A,F,H[ ]A,B[ ]A,F,G[
]8[ ]13[ ]14[ ]9 [
]A,F,G,I,E[ ]A,F,H[ ]A,B[ ]A,F,G,I[
]A,F,G,I,J[ ]8[
]10[ ]15[ ]13[ ]14[
]A,F,G,I,J[
]10[
Important Notes
• A* Algorithm is one of the best finding algorithms.
• But it does not produce the shortest path always.
• This is because it heavily depends on heuristics.
