pwscripts

Scripts supporting the design, execution, and analysis of Polyworld simulations

Prerequisites

• Polyworld
• Java 8 with Ant
• Python 3 with NumPy and pandas

Installation

1. Clone the repository:

   $ git clone https://github.com/smailliwcs/pwscripts.git
   

2. Initialize and update submodules:

   pwscripts$ git submodule init
   pwscripts$ git submodule update
   

3. Build the JIDT submodule:

   pwscripts/java/lib/jidt$ ant
   

4. Build the Java code:

   pwscripts/java$ ant
   

5. Configure environment variables:

   • PATH: Add pwscripts/shell.
   • CLASSPATH: Add pwscripts/java/lib/jidt/infodynamics.jar and pwscripts/java/bin.
   • PYTHONPATH: Add polyworld/scripts.

Usage

Replacing Polyworld's default (rotating oblique) view with a static overhead view

polyworld$ ln -s ../pwscripts/config/static.mf gui.mf

Running a "legacy" simulation

polyworld$ ./Polyworld ../pwscripts/config/legacy.wf

Running a "modern" simulation

polyworld$ ./Polyworld $(../pwscripts/config/modern)

To grow "poisonous" food items, pass the -D poison option to modern.

Running driven simulations

polyworld$ pwdriven ../pwscripts/config/legacy.wf 10 runs/driven

To ensure simulations run to completion, pass the --complete option to pwdriven. This discards simulations that terminate prematurely (e.g., due to a population crash).

Running passive simulations

polyworld$ pwpassive runs/passive runs/driven/*

Encoding a Polyworld "movie" in another video format (requires FFmpeg)

The entire simulation:

polyworld$ pmvencode run/movie.pmv run/movie.mp4 640 480

A segment of the simulation:

polyworld$ pmvencode run/movie.pmv run/movie.mp4 640 480 -- -s 9001 -e 10000

Calculating genetic diversity

polyworld$ ./bin/genetics runs/legacy | java Diversity 0 > diversity.txt

Calculating information-theoretic metrics

Post hoc:

polyworld$ ./bin/timeseries in-vitro runs/legacy death 1 10 1000 | java Complexity > complexity.txt

On the fly:

polyworld$ mkfifo timeseries
polyworld$ java Complexity < timeseries > complexity.txt &
polyworld$ ./Polyworld ../pwscripts/config/legacy.wf --TimeSeriesLog '{ On True; }'
polyworld$ rm timeseries

Information-theoretic metrics include Complexity, Entropy, CompleteTransferEntropy, CollectiveTransferEntropy, and InfoDynamics (which calculates the following metrics in one pass: active information storage, apparent transfer entropy, and separable information). In general, multiple and/or longer time series may be required to achieve numerical stability. To accomplish this when calculating post hoc, increase the first and/or last numeric arguments to timeseries. When calculating on the fly, specify the Repeats and/or Steps subparameters of TimeSeriesLog. For some metrics, additional arguments to java are required.

Calculating other metrics

Refer to the Python code's inline documentation:

pwscripts/python$ python calculate.py --help