Complete.Org: Mailing Lists: Archives: freeciv-ai: June 2003:
[freeciv-ai] Re: Idea for a AI testcase

[freeciv-ai] Re: Idea for a AI testcase

[Top] [All Lists]

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index] [Thread Index]
To: freeciv-ai@xxxxxxxxxxx
Subject: [freeciv-ai] Re: Idea for a AI testcase
From: "Per I. Mathisen" <per@xxxxxxxxxxx>
Date: Sun, 15 Jun 2003 11:52:04 +0000 (GMT)

On Sun, 15 Jun 2003, Raimar Falke wrote:
> A big island and two players. Each player has a city and an equal
> number of units of the same type. These are places like this.
>      uuu                                 UUU
> c    uuu                                 UUU   C
>      uuu                               UUU
> Both players know the whole map. The island has random terrain. The
> cities can't produce new units. Units can't be upgraded.

Does the city have a barracks? Are the units supported by the city? If the
latter, then the city is the victory condition. Otherwise everything must
be killed (no strategy involved, just small-scale tactics).

> Good actions by the AI:
>  - move unit to a tile with a high defence value
>  - fortify unit
>  -move unit to a tile with a high defence value and fortify unit
>  - move unit into city
>  - move unit into city to heal
>  - planned attack (multiple units vs one enemy unit)

Units should fortify correctly, and I don't remember if we fixed all the
new don't-fortify-correct bugs that were introduced. Also going to a
barracks to heal is a good idea, and we do that. We defend cities.

We do not exploit defensive terrain. Recognizing defensive positions is a
hard problem.

But you forgot one important action: Stopping out of range of enemy units
when you do not have enough movement to attack it outright. We do this,
but we are not good at breaking a deadlock of units both doing this.

  - Per

[Prev in Thread] Current Thread [Next in Thread]