Adding nodes to PointGraph

General
1

Hello i am interested is there a way to add points to a pointGraph and how to go about it…the updateGraph doesnt seem to work.and scan is too costly.
i want to do it runtime,so performance is important.

2

See the Documentation here: http://arongranberg.com/astar/docs/graph-updates.php

If you have any further question feel free to ask!!

If performance is an issue, do you have to use a pointGraph? There are some better options for dynamic Graphs including:

(Free) Grid-Graph updating, (Pro only) Subtractive Recasting, or (Pro only) RVO Based redirection

Grid-Graph’s just set nodes to walkable or unwalkable making things VERY easy,

Subtractive Recasting Requires a bit of forethought, but if you have set locations that are dynamic (Like destructible walls and such) you can PreCash all the variations (Note: this is very unwieldy in heavily dynamic games)

RVO Based redirection is super easy, and takes almost no performance: But can give sub optimal paths when there are large obstacles~ This is the same system the game SC2 uses for avoiding buildings and large units for example

3
Subtractive Recasting Requires a bit of forethought, but if you have set locations that are dynamic (Like destructible walls and such) you can PreCash all the variations (Note: this is very unwieldy in heavily dynamic games)
A side note: this has been improved in the latest beta: http://www.arongranberg.com/2013/08/navmesh-cutting/

Adding points directly to a point graph will be made possible, it’s not quite there yet, but will be soon.

4

A side, side note: Could you use recast cutting to remove obstacles? (IE: I blow up a wall with a rocket launcher)

If not, is there some way the space occupied by the obstacle could be made walkable?

5

Nope, not possible.
It is also possible in the beta to recalculate whole tiles (small regions of the recast mesh) to accomplish that, but it is much slower, not really real time. Depending on tile size, it takes maybe 100 ms (rough estimate, haven’t actually tested), compared to cutting, which takes from 2 to 20 ms in my test cases.