Agentic Evolution: Cooperators vs Exploiters

Toy agent-based model of frequency-dependent selection: C cooperates, P exploits (psychopathy-like strategy proxy).

Population

Init P%

Benefit (b)

Cost (c)

Exploit take (e)

Policing prob (p)

Punishment size

Mutation rate

Interaction radius

Run speed (ticks/sec)

Max ticks/frame

Partner choice (avoid low rep)

Deterministic seed

Use fixed RNG

Seed value

Tick0

P frequency0.000

Mean energy0.00

Mean rep0.00

Punishments/tick0

State consistencyOK

If P gets common, cooperation and reputation fall and punishment rises; that can push P back down. This is the agent-based analogue of frequency-dependent selection.

World

Green=C, Pink=P, Yellow=punished

Toroidal grid

Local interactions + replacement

Cooperator (C) Exploiter (P) Punished

Time series

P frequency (pink) and mean energy (blue)

Rolling window

Last ~1200 ticks