Re: pathfinding refueling

[Eric McDonald <mcdonald at phy dot cmich dot edu>]

On Wed, 17 Dec 2003, Peter Garrone wrote:

> I propose that only one material per unit type be considered for
> refueling control during pathfinding.

Why only one? In the case of Bellum, c is every bit as important 
as f1 or f2. A unit will become paralyzed without c.

> The pathfinding is helper-ai, not "refereed" code, so we try to keep it out
> of the GDL. 

'material-to-move' and 'consumption-per-move' are part of GDL. I 
see no reason why the pathfinding should not pay attention to 
those very important tables, even if the referee must as well.

> 1) The storage of that material is greater than zero

Does this include treasury storage?

> 2) The consumption per move, or base consumption, is greater than zero

Fair enough.

> 3) The distance that the unit can receive the material is zero.

How is this a relevant criterion?

> I cannot see any units that require more than one type of fuel. In
> bellum for example, the "c" material is not a fuel because it has a
> send/receive distance greater than zero. 

A unit is not guaranteed to receive more "c" just because the 
'in-length'/'out-length' happens to be > 0. "c" is just as 
relevant as "f1", IMO. Any refueling path heuristic should not 
assume some hypothetical supply of a material; all calculations 
should be done with material in hand.

> The main reason only one material can be a fuel are as follows:
> - To be within range, a final untaken trip is calculated to a supply
>   point, and having multiple supply points makes it all too difficult.

One can always choose to be greedy, meaning that when a unit stops 
to resupply one material, it also tries to grab any others 
relevant to its journey.

Also, if the journey is sufficiently long, then one must indeed 
have multiple supply points, even if for one material.

> - It would require a variable length array to be added to pathfinding
>   structures, making it a bit cumbersome.

I hope this doesn't sound too harsh, but I would prefer 
correctness over not doing something because it is a "bit 
cumbersome".

> task. So an aircraft carrier could be assigned a refueling task to
> support an invasion. It just sits there doing nothing while aircraft
> automatically go to it, refuel, get ammo, and return to the attack.

OK. That is somewhat appealing.

> I would add some code to take ammo as well if refueling.

Sure, might as well.

  Regards,
    Eric