OJ Board

What's a good way to hash? I use STL's <set>, but gets TLE. Should I write my own hash class?
I also use STL's<set>, but it is not enough; notice this
2. heuristic to estimate if a knight will meet other knight on the board, if it won't meet other, just add the number of it's reachable sqaure ...

.. wrote:My idea is first dump out all the reachable squares by each knight, then use sorting to deduplicate. You can use other method to deduplicate.
Sorting is not fastet method, but it is quite easy to code.

May i see you code? I can not imagine yours is so fast!
Please send to wiltord@eyou.com, thanks!

I don't know how you can come up with 100 000000 such a big number.
Consider we have 50 knights where every one can walk 50 steps.
Maximum number of covered square < 50 x 35000 = 1750000
Even you sort all the 1750000 squares in a qsort, the number of operation should be < 1 0000000 (although this ...

I don't know how you can come up with 100 000000 such a big number.
Consider we have 50 knights where every one can walk 50 steps.
Maximum number of covered square < 50 x 35000 = 1750000
Even you sort all the 1750000 squares in a qsort, the number of operation should be < 1 0000000 (although this ...

I don't know how to implement the trie, But get AC with divide and conquer algorithm. Divide the string set to two parts, then gets the dominant strings in the two parts respectively with recursion, i call the results as candidate dominant strings of part1 and candidate dominant strings of part2 ...

A is in the ready queue....A then executes an unlock statement.

Hi, junbin. I think you don't really understand the states of a program.
In this problem, a program may be in three states, running(being processed by cpu),waiting(in the ready queue),blocked(in the blocked queue). So,if A is in the ...

I' got AC :lol: ;following is my approach:

First,use BFS to know whether there is a path from source to destination.If it exists,record the length,and if it is prime,then that is the result,otherwise it is the min composite number,in this case, we shall goto IDS procedure.

Then,I use IDS to search ...

sohel wrote: I used IDS.

Do you use 2,3,5,7 as the threshold and simplily do IDS?

I also tried BFS,but still tle. Anything else I should do?

who can help me? I can't really find an effective way to avoid tle. thanks before.

I run a IDS (Iterative Deepening Search) keeping track of the cells I visit while searching within a certain depth. If this track is kept, then we are not going to explore the cells in that depth again if it leads to unsuccessful searches
please tell me more specificly,how to do with the keeping ...

I have found the way now,it is really i misunderstood the move rules.thank you! [/quote][/cpp]

how to understand the second data set in the sample. i can't find a two steps move from (0,0) to (4,4).

I have found the so called sophisticated structure in <<introduction to algorithms>>,open my eyes.

Alexander Grushetsky wrote: And for each step I use Dijkstra with complexity E*logV.

how could Dijkstra be complexity of E*logV?