Pathing - return path divergance

[Guest]

Archived from groups: comp.ai.games (More info?)

Begin at Start, travel to a waypoint, and return to start. The path TO the
waypoint and the path FROM the waypoint should be the same path.

So far, I've tried RandomBounce, SimpleTrace, RobustTrace, BreadthFirst,
DepthFirst, IDDF, Dijkstra's, BestFirst, A*, and IDA*.

Each of those algorithms produces a return path that is a rotated image of
the original. Thus, the course travelled ends up being two different paths,
which is not good for me.

Has anyone seen work on this?

 

[Guest]

Archived from groups: comp.ai.games (More info?)

AngleWyrm wrote:

> Each of those algorithms produces a return path that is a rotated image of
> the original. Thus, the course travelled ends up being two different paths,
> which is not good for me.

Is there any reason why you don't calculate the return path by reversing
the path from start to the waypoint?

If there is such a reason, then I suggest that you take a look at the
tie-breaking of your algorithm; I would guess that this causes the
mirroring.

--
mail1dotstofanetdotdk

 

[Guest]

Archived from groups: comp.ai.games (More info?)

"Bjorn Reese" <breese@see.signature> wrote in message
news:4204fbaa$0$13775$ba624c82@nntp03.dk.telia.net...
> AngleWyrm wrote:
>
> > Each of those algorithms produces a return path that is a rotated image
of
> > the original. Thus, the course travelled ends up being two different
paths,
> > which is not good for me.
>
> Is there any reason why you don't calculate the return path by reversing
> the path from start to the waypoint?
>
> If there is such a reason, then I suggest that you take a look at the
> tie-breaking of your algorithm; I would guess that this causes the
> mirroring.

Yes, and thank you; mirroring is exactly what I was looking for (as opposed
to rotation). The part where the next location is picked was doing it always
in one direction (clockwise), and so now I have it doing the picking
counterclockwise on return trips. Thanks again.