This shows you the differences between two versions of the page.
| — |
guide:a-star [2009/01/09 16:32] (current) 199.4.155.10 created |
||
|---|---|---|---|
| Line 1: | Line 1: | ||
| + | Here's an idea regarding the A* algorithm. | ||
| + | > One interesting feature of all these grid search methods, is that they | ||
| + | > have not only the current best path to the goal, but also have *a* | ||
| + | > path, if not the *best* path, from every point searched, back to the | ||
| + | > origin. So, for example, rather than several enemies finding their own | ||
| + | > path to the player, the program can instead find the path from the | ||
| + | > player to the most distant enemy, also noting which *other* enemies | ||
| + | > are on the paths searched. Any enemies not found can be searched for, | ||
| + | > also recording incidental enemies found. This puts to use what would | ||
| + | > otherwise be wasted search time and data. The result being that a | ||
| + | > route will have been found from every enemy back to the player. | ||
| + | > | ||
| + | > This seems like it might be useful. For example, creating a search | ||
| + | > path matrix over terrain, using the levelness and down-hillness of the | ||
| + | > terrain as the heuristic, and searching from the lowest point to the | ||
| + | > highest point, might create a map of the down-hill flow over the | ||
| + | > terrain toward the lowest point. Or something. | ||
| + | > | ||
| + | > James Steiner | ||