3 - Informed Search Flashcards
Informed Search Techniques …
… have additional information about more promising paths and are therefore more efficient than uninformed search algorithms.
The choice of the next node is based on ________.
The choice of the next node is based on THE EVALUATION FUNCTION f(n).
A heuristic h(n) is admissible iff …
UNDERESTIMATE: h(A) <= c(A; goal)
A heuristic h(n) is consistent iff …
TRIANGLE INEQUALITY: h(A) <= c(A;B) + h(B)
The performance of the heuristic depends on the ________.
… on the QUALITY OF THE HEURISTIC.
What are good possibilities to obtain decent heuristics?
Relaxed problems and pattern databases
Greedy Best-First Search (GBFS)
Evaluation function is heuristic: f(n) = h(n)
GBFS explores ________ which is ________.
GBFS explores THE NODE WITH THE LOWEST HEURISTIC FUNCTION which is NOT ALWAYS OPTIMAL.
GBFS: Completeness
GBFS is complete if graph search is used
GBFS: Optimality
GBFS is NOT optimal
GBFS: Time/Space Complexity
GBFS has EXPONENTIAL time complexity O(b^m) and EXPONENTIAL space complexity O(b^m)
A* Search (A*)
Evaluation function is sum of heuristic and path cost: f(n) = h(n) + c
A*: Completeness
A* is complete if costs > 0
A*: Optimality
A* is optimal iff costs > 0 && heuristic admissible
A*: Time/Space Complexity
A* has EXPONENTIAL time complexity O(b^(eps*d)) and EXPONENTIAL space complexity O(b^(eps*d))