0

I am studying Iterative Deepening from this link. My main concern is with Overhead. That link says that

The higher the branching factor, the lower the overhead of repeatedly expanded states


There is no explanation given for this statement and also no convincing arguments are given at that link. I am searching the reason behind this statement because I think that overhead should increase as branching factor is increasing and it also means that no of nodes are increasing then how overhead is reducing?

Till now I did not find anything reasonable and helpful. If someone can help in correcting my concepts then I would be thankful to you.

Hammad Hassan
  • 1,192
  • 17
  • 29

1 Answers1

0

The answer to your question is in the formula above the statement 

 (d)b + (d-1)b^{2} + \cdots + 3b^{d-2} + 2b^{d-1} + b^{d}

The cost of doing a BFS - which is to what you should compare is -- 

 b + b^{2} + \cdots + b^{d-2} + b^{d-1} + b^{d}

Hence the overhead is 

   (d-1)b + (d-2)b^{2} + \cdots + 2b^{d-2} + 1b^{d-1}

Obviously this is heavily influence by the branching factor. especially when looking at the last term 1b^{d-1}

CAFEBABE
  • 3,983
  • 1
  • 19
  • 38
  • You said, "Obviously this is heavily influence by the branching factor", does it means overhead increases with increase in branching factor? – Hammad Hassan Mar 14 '16 at 08:42
  • yes if the solution is in the second level of the search it is grow quadratic, third level cubic,.... This is kind of disastrous, however, the algorithms run time runs `b-times` that fast, hence, the overhead is clearly dominated – CAFEBABE Mar 14 '16 at 17:51