15.08.2013 Views

General Computer Science 320201 GenCS I & II Lecture ... - Kwarc

General Computer Science 320201 GenCS I & II Lecture ... - Kwarc

General Computer Science 320201 GenCS I & II Lecture ... - Kwarc

SHOW MORE
SHOW LESS

Create successful ePaper yourself

Turn your PDF publications into a flip-book with our unique Google optimized e-Paper software.

c○: Michael Kohlhase 466<br />

Iterative deepening search: Properties<br />

Complete Yes<br />

Time (d + 1)b 0 + db 1 + (d − 1)b 2 + . . . + b d ∈ O(b d+1 )<br />

Space O(bd)<br />

Optimal Yes (if step cost = 1)<br />

(Depth-First) Iterative-Deepening Search often used in practice for search spaces of large,<br />

infinite, or unknown depth.<br />

Comparison:<br />

Criterion<br />

Breadthfirst<br />

Uniformcost<br />

Depthfirst<br />

Iterative<br />

deepening<br />

Complete? Yes ∗ Yes ∗ No Yes<br />

Time b d+1 ≈ b d b m b d<br />

Space b d+1 ≈ b d bm bd<br />

Optimal? Yes ∗ Yes No Yes<br />

c○: Michael Kohlhase 467<br />

Note: To find a solution (at depth d) we have to search the whole tree up to d. Of course since we<br />

do not save the search state, we have to re-compute the upper part of the tree for the next level.<br />

This seems like a great waste of resources at first, however, iterative deepening search tries to be<br />

complete without the space penalties.<br />

However, the space complexity is as good as depth-first search, since we are using depth-first<br />

search along the way. Like in breadth-first search, the whole tree on level d (of optimal solution)<br />

is explored, so optimality is inherited from there. Like breadth-first search, one can modify this<br />

to incorporate uniform cost search.<br />

As a consequence, variants of iterative deepening search are the method of choice if we do not<br />

have additional information.<br />

Comparison<br />

251

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!