We have talked a bit about units, little creatures that interact in a computer game or educational simulation.
While most simulations have broad categories of units, increasingly when describing constituents, be they customers or voters, the more accurate simulations use at least ten or fifteen variables to capture attributes such as spending ability, preferred/trusted media, relationships with other units, and interests and concerns.
Then, when it comes to such questions as "will they buy from you or a competitor?" "for whom will they vote?" or "are they happy?" the sim does the necessary fuzzy logic to come up with a discreet answer.
In physics, we have wrestled with the question of: is something a wave or a particle?
Simulation designers will likewise wrestle with the question, is behavior best modeled by an abstract high level system, or by creating individual units.
Saturday, January 21
Subscribe to:
Post Comments (Atom)
2 comments:
Simulations are fantastic for testing models and theories, tweaks in the systems, variable, emergent behavior, etc.
Are there any functional "predictive engines" out there that can, besides model behavior, predict outcomes? Or is that in the snakeoil category?
I think yes, but the real question is how predictive.
Post a Comment