Re: Revised exploring_worth function

[Hans Ronne <hronne at pp dot sbbs dot se>]

>Maybe I am misunderstanding the meaning of "explorer worth."  Does
>it include factors such as building costs (particularly the number
>of turns required to build one), visibility range (and whether it,
>in turn, can be seen -- e.g. subs), and defensive capability?

Sure. The building cost is particularly important. And other things, such
as where the unit can go, should also factor in. An air unit should be more
worth than land or sea units. Since most of the interesting stuff to be
found usually is on land, one could also argue that land units should be
more worth than sea units.

The current AI worth functions are just a first stab at providing
evaluations that make sense.

>Maybe this is becoming too game-specific, but would it be possible
>to divide the (known) world into theaters, and then have there be
>exploration/reconnaissance plans for each theater, with the producing
>units in each theater set up to produce what is needed for those
>plans?  That might help to avoid duplicate effort (e.g. if each town
>tried to produce local reconnaissance units, even in a densely
>populated area).  How to effectively determine theaters is left as
>an exercise for the reader.  B-)

This is exactly how the AI code works. It has a drawback, though. The AI:s
forces easily get fragmented because of this. So it is very important that
units are easily transferred between theaters. This is something I
addressed in the latest series of AI bug fixes.

Hans

Hans Ronne

hronne@pp.sbbs.se