Friam Digest, Vol 38, Issue 3

Posted by Marcus G. Daniels-2 on Aug 06, 2006; 3:30am
URL: http://friam.383.s1.nabble.com/Friam-Digest-Vol-38-Issue-3-tp522318p522334.html

Robert Cordingley wrote:
> Unfortnately, neither business management nor governing is a total
> disclosure game. [..]  I wonder, what hope is there of computationally
> solving problems involving millions of agents in dozens of countries
> acting in myriads of ways (for example)?  May be that wasn't the question.
>  
Well, one motivation for a computational model is to get ideas about
what needs to be measured in order to make useful predictions, but not
necessarily to be the mechanism of prediction.  Computationally, making
useful predictions could be as simple as a regression once a set of
appropriate signals have been acquired, e.g. by measuring dynamics in a
simplified simulated world and finding the same dynamics in the real
world.  Running a simulation of millions of simple agents with thousands
of variant scenarios ought to be doable for a government or big company,
but even that following that approach doesn't mean anyone is actually
thinking in terms of `solving' the game, or even claiming to know the
rules.  Rather, the goal is just to shift the odds.

Marcus