Post date: Aug 20, 2008 11:25:4 PM
In ATI's Froblin demo they maintain separate influence maps for each goal. Although the application is different it occurs to me that this might be easier to work with than a unified influence map, what I am currently calling a potential field. So separate out the excitatory and inhibitory components of the shunting equation and store them in positive and negative influence maps respectively, dunno, might be worth looking into.