Risk sensitivity and assortment in social dilemmas

Uncertainty and risk are key features of many social dilemmas, where individual decisions are often made with imperfect knowledge and variance in outcomes. In this paper, we investigate the evolutionary dynamics of the well-known public goods game using a recently introduced modelling framework encapsulating risk-sensitive assortment. Here, the population of mobile agents playing the game is divided into fixed-sized interaction groups. Individuals are defined by a single genetic trait—a risk sensitivity trait—that guides their decision-making. This trait is mapped to a continuous range of investment levels and also provides a mechanism to guide mobility (migration) decisions. Detailed computational simulation experiments confirm the relationship between risk orientation, decision-making and mobility in the game. As the size of each group increases, assortment levels tend to decrease and risk-averse individuals tend to dominate the population. However, in many scenarios, there was high variance in the proportion of ‘cooperators’ both in groups and between different groups, suggesting that risk-seeking behaviour is an emergent property of mobility induced positive assortment.