Evolutionary Game Theory: Agent-Based Simulations in R

This directory contains R scripts that run agent-based evolutionary game theory simulations. The basic structure of the simulations is as follows:

Simulations are run over discrete "time periods".
In each time period, agents are randomly partitioned into pairs.
Each pair of agents plays a cooperate-defect style game (similar to, but not necessarily the same as the prisoner's dilemma), whose outcome determines each agent's chance of surviving to the next time period.
- Each agent's "play" (decision to cooperate or defect) depends on their and their opponent's 2-period history of cooperate/defect decisions, together with that agent's "genetic strategy".
  - A genetic strategy is a binary vector that determines, in each of the 2^(2+2) = 16 possible 2-period histories for a pair of agents, what an agent's decision will be.
  - I use the term "genetic strategy" because this length-16 binary vector can be mutated, i.e. when an agent self-replicates there might be an error copying any of his 16 binary genes (so that 0000000000000000 might be replicated as 0010000000000000).
- In this project, the following protocol is used:
  - 0 = cooperate
  - 1 = defect
Some agents die, according to their survival probability.
Some of the surviving agents reproduce, i.e. self-replicate.
When agents self-replicate, each of their 16 binary genes has a chance of being 'mutated', i.e. changed from 0 to 1 or vice-versa.

The inspiration for this project, and for many parts of the model used, comes from this paper by Kristian Lindgren, which I initially discovered through this working paper by W. Brian Arthur.

Setup

To get the scripts running in R, you must:

Install R.
Download the game_flexible_popn.R and play_flexible_popn.R scripts from this repository.

In R, source these two scripts:

On Windows, something like

source("C:\\Users\\yourusername\\Desktop\\game_flexible_popn.R")
source("C:\\Users\\yourusername\\Desktop\\play_flexible_popn.R")

On Mac, something like

source("/Users/yourusername/Desktop/game_flexible_popn.R")
source("/Users/yourusername/Desktop/play_flexible_popn.R")

Then call the game_flexible_popn() function:

game_stats <- game_flexible_popn(T = 1000)