NetLogo modelling tool
NetLogo is an interesting tool for modelling dynamic systems. It's written in Java and has an extensive library of pre-built models. Both the tool and the library are downloadable.
The one I started with is this one, which models wealth distribution in a simplified world.
Also this one, which models prisoners' dilemma under the following circumstances:
Here's the payoff matrix:
Cooperate Defect
Cooperate (1, 1) (0, alpha)
Defect (alpha, 0) (0, 0)
The world starts out with some percentage of agents cooperating on their first turn. On the second turn each agent cooperates or defects depending on what the majority of its neighbors have done.
The payoff for cooperating is 1. The payoff for defecting is a variable, and always greater than the payoff for cooperating. But if someone defects, it's more likely that the others around them will defect also, hurting them in the end. So with a relatively low payoff for defecting, the simulation shows cooperation spreading through the population.
As the defection award grows the spread of cooperation slows. When the rate is around 1.2x you get tiny islands that never learn to cooperate. Around 1.4x the islands are bigger. At around 1.7x selfishness spreads rampantly and surrounded islands of cooperation are ultimately wiped out.
The one I started with is this one, which models wealth distribution in a simplified world.
Also this one, which models prisoners' dilemma under the following circumstances:
Here's the payoff matrix:
Cooperate Defect
Cooperate (1, 1) (0, alpha)
Defect (alpha, 0) (0, 0)
The world starts out with some percentage of agents cooperating on their first turn. On the second turn each agent cooperates or defects depending on what the majority of its neighbors have done.
The payoff for cooperating is 1. The payoff for defecting is a variable, and always greater than the payoff for cooperating. But if someone defects, it's more likely that the others around them will defect also, hurting them in the end. So with a relatively low payoff for defecting, the simulation shows cooperation spreading through the population.
As the defection award grows the spread of cooperation slows. When the rate is around 1.2x you get tiny islands that never learn to cooperate. Around 1.4x the islands are bigger. At around 1.7x selfishness spreads rampantly and surrounded islands of cooperation are ultimately wiped out.
Comments
Post a Comment