Abstract
Interactions in the real world are subject to mistakes and miscommunications. The presence of this noise challenges cooperation, as one party cannot determine whether the other party did not cooperate on purpose. Prisoner’s dilemma strategies like Tit-for-Tat (TFT) perform badly once noise is present. Recent studies that harsh environments promote cooperation do not take noise into account. We show that the adversity of the environment benefits cooperators and can make cooperation more robust against mistakes. Harsher environments also encourage greater generosity to cope with noise. Yet when uncertainty is substantial due to higher probability of mistakes or more potential defectors in the environment, contrite behaviours are the most successful.
Original language | English |
---|---|
Pages (from-to) | 414-415 |
Number of pages | 2 |
Journal | Belgian/Netherlands Artificial Intelligence Conference |
Publication status | Published - 2020 |
Event | 32nd Benelux Conference on Artificial Intelligence, BNAIC 2020 and 29th Annual Belgian-Dutch Conference on Machine Learning, BeneLearn 2020 - Virtual, Online, Netherlands Duration: 19 Nov 2020 → 20 Nov 2020 |