Game Development Reference
In-Depth Information
Figure 18.5
Multiple agents trying to reach a single node.
each other
s comfort zones. This ended up working really well and solved most of
our issues concerning people running into each other.
Having multiple agents all moving around using complex pathing graphs can be very
taxing on the system. In larger game worlds, a common practice is to allow the A*
algorithm to stop at any time so that a single path can be built across multiple
frames. This is easy enough to implement with the system you
'
ve built. All you
need to do is to store the
AStar
object for each path being built and have an event
sent when the path is done. This sounds like a perfect job for a
Process
object. In
Chapter 20,
'
“
Introduction to Multiprogramming,
”
you
'
ll learn an even better solution
using threads.
Here is a short list of topics I
'
ve found very helpful
in becoming a better AI
programmer:
n
Artificial Intelligence for Games, Ian Millington, published by The Morgan
Kaufmann Series in Interactive 3D Technology
