20231120-MTOA: Single-tasking agent populations cannot achieve equitable task exploration.

This archive contains the results of a multi-agent simulation experiment [1] carried out with Lazy lavender [2] environment.

Experiment Label: 20231120-MTOA

Experiment design: Single-tasking agents are periodically replaced by new, naive ones. The new agents are trained by previously existing agents. The selection of the teaching agents is based on different selection criteria. These include: (a) low/high acquired compensation, (b) low/high success rate, (c) random selection.

Hypotheses: Favoring the reproduction of agents with the lowest success rate will improve the maximum accuracy and re-balance the tasks representation.

Detailed information can be found in index.html or notebook.ipynb.

[1] https://sake.re/20231120-MTOA
[2] https://gitlab.inria.fr/moex/lazylav/