Complete.Org: Mailing Lists: Archives: freeciv-ai: April 2002:
[freeciv-ai] Re: [Freeciv-Dev] Re: [RFC][Patch] AI can fly v2 (fwd)

[freeciv-ai] Re: [Freeciv-Dev] Re: [RFC][Patch] AI can fly v2 (fwd)

[Top] [All Lists]

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index] [Thread Index]
To: "Ross W. Wetmore" <rwetmore@xxxxxxxxxxxx>
Cc: per@xxxxxxxxxxx, freeciv-ai@xxxxxxxxxxx
Subject: [freeciv-ai] Re: [Freeciv-Dev] Re: [RFC][Patch] AI can fly v2 (fwd)
From: Raahul Kumar <raahul_da_man@xxxxxxxxx>
Date: Fri, 12 Apr 2002 18:56:48 -0700 (PDT)

--- "Ross W. Wetmore" <rwetmore@xxxxxxxxxxxx> wrote:
> >I suggest we put in analyzing/profiling code. Unfortunately, I can see big
> >slowdowns happening because of all this extra code. No way to avoid it.
> Why is AI so incredibly computationally expensive!
> Because people get hung up on fuzzy concepts like "cheat" without thinking
> about or understanding the "practical" implications and thus insist on 
> rebuilding the human mental processes in code.
> To build a fuzz map, you can start from a real map and fuzz it, or 
> start from zero and try to build it by (expensive) collection of 
> partial real data.

This is where my problem lies. Raimar has spoken. The server side AI is doomed.
I'd like to experiment with a server side AI that does *NOT* start with the
real full map. This fuzz factor is bad for reasons I already mentioned in
previous mails. It's just too easy to play with 0 fuzz factor. We're only
human, the code is willing, but the programmer is weak.

> >> I think it is just some calculations that are screwed up somewhere. Don't
> >> know where, though.
> >
> >Guess. Anything at all would be handy. I've looked at f_s_t_k, and
> >find_beachhead, and I suspect that one of these has the bug. 
> Raimar refuses to fix the find_a_beachhead bug. He remove the fix from
> an earlier patch and rejected three separate people's rediscovery of
> this over several months, R_T_F_L :-).

I missed this. Raimar, why did you reject the fixes? Which one was most
acceptable to you, or do you want a totally different fix?

Be specific.

> Add a city value to city_reinforcements_cost_and_value() to fix f_s_t_k:
>   /* Add a basic city cost, i.e. SETTLERS + 20 per extra pop
>    * This insures that even empty cities have value.
>   */
>   pcity->ai.f += (pcity->size + 1) * 20; 

Syela specifically removed this. I read the comments in the code, and it seems
Syela wanted empty cities to have less value. Why - who the hell knows. Can I
ask you to send this in as a patch? I'm already responsible for Per's
portattacks 3, I don't want to be the person who takes over other people's
That's what we have maintainers for.

> You probably want to change the g to 1 in kill_desire() calls for cities
> in advMilitary.c and aiunit.c to reduce the desirablility of attacking 
> heavily defended cities over weakly defended ones. It is not a complete 
> fix, but it helps.

Don't hold back. Inquiring minds want to know. What is a complete fix?


Of course, it is very important to be sober when you take an exam. Many
worthwhile careers in the street-cleansing, fruit-picking, and
subway-guitar-playing industries have been founded on a lack of understanding
of this simple fact.
{Moving Pictures, 1990, Terry Pratchett}

Do You Yahoo!?
Yahoo! Tax Center - online filing with TurboTax

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