Geek Answers Handbook

The optimal algorithm for finding paths in a matrix that does not fully fit into memory

I ran into a serious problem:

Imagine that I have a map of the whole country, represented by a huge matrix of cells. Each cell represents 1 square meter of territory. Each cell is represented as a value doublebetween 0 and 1, which represents the cost of moving the cell.

Obviously, the card does not fit in memory.

I am trying to combine my mind to calculate the optimal path for the robot, from the starting point to the end position. The first idea that I chose was to create a moving window like TCP, with a minimal map of a real map around the moving robot and running the A * algorithm inside, but I have to deal with some problems with maps with huge walls, bad pathfinding, etc.

I am looking for literature on A * -like algorithms, and I could not imagine approximating what would be a good solution to this problem.

I am wondering if anyone has encountered a similar problem or can help with the idea of ​​a possible solution!

Thanks in advance:)

+5
source share
3

, , :

  • . () . , : , . , , : A B. . , .

  • , - . , , .

  • ( ) . , : -, -, , , , . , , , , .

+1

- , , / , -? ? ( , , .) - ?

, - 100 100 100\times 100 . ( 1 ), Dijkstra, 100\times 100 square.

, Dijkstra ? I.e., , .

, ?

+1

,

- . 31 , 60 . ,

(x1, x2, x3..... x60)

- 60 . . - . - monte carlo,

http://en.wikipedia.org/wiki/Parallel_tempering

, , . , , . Newtons (, , BFGS), .

http://en.wikipedia.org/wiki/Local_minimum

http://en.wikipedia.org/wiki/BFGS

. , - " ".

:

  • , , ?
  • ?
  • Can the robot make sharp turns or do you need additional physical modeling to improve the cost function?
+1
source

More articles:


All Articles