C[omp]ute

From: "andrew cooke" <andrew@...>

Date: Fri, 19 Jan 2007 12:57:13 -0300 (CLST)

After experimenting with this simulation -
http://www.bbc.co.uk/sn/hottopics/climatechange/climate_challenge/ - I
wondered if there was any work that tried to find good heuristics for
optimising financial models.

You could imagine generating random "non-parametric" heuristics (always
keep water supply above X) and running them against a complex model in the
hope that there's some robust combination that are simple enough to act as
an effective guide for politicians.

The simulation at the BBC site is rather frustrating because (as far as I
could tell) it gives no feedback about the financial or political state of
the economy.  So I ended up optimising the system by the indicators given
(ecological) only to find, on the "results" page that i had ruined the
economy and encouraged nationalistic politics (this despite remaining in
democratic power for 80 years!)

Andrew