I want to try to implement HPA*, which in the paper "Near Optimal Hierarchical Path-Finding by A.Botea,M Muller, and J.Schaeffer. http://www.cs.ualberta.ca/~jonathan/Papers/Papers/jogd.pdf However I have difficulty in understanding it. May be partly because my first language is not English. The part that confuse me is the Table 1 results (page 25) for the example results that is given. I just can not figure out how to get the number of expanded nodes in L-1 and L-2 rows. The explanation in the paper is on page 12-15, however when I read it just make me confuse. Can someone please help to understand it? Thank you

"I just can not figure out how to get the number of expanded nodes in L-1 and L-2 rows."

They most likely count the number of unique nodes that have been put on A*'s open list during the A* searches for level 1 respectively 2.

William

First, Thank you for the response.

That's is what I think too.
However on the table1 the L-1 row for the SG part, it has 16 nodes. When I try to calculate it by my own it is 17. Did I make a mistake ?

One more question, for the low-level search, the heuristic that the paper used is Manhattan distance. How about for the level 1 and 2 abstraction ?

Thank you