Risk Sensitivity Emergence in a Prey-Predators Agent-based Model

Risk sensibility is a phenomenon that appears in different field like psychology, economics and biology. Although diverse in several features, all these disciplines assume that risk-sensible individuals take decision characterized by different level of caution under conditions of uncertainty. This feature could emerge from an evolutionary process in certain circumstances. Nevertheless, there is a lack of studies about the effect of the environmental characteristics on the evolutionary emergence of risk-sensible behaviour in a population.
To address this issue, we developed an evolutionary prey-predator agent-based model. In this model, a prey dies when a predator reaches it or when it runs out of energy. The longer a prey stays alive, the higher is its expected number of offsprings. Preys have limited information available; they can detect energy sources and predators only in the vicinity. When a prey perceives an energy source and a predator, it decides where to head computing a utility function whose parameters are its genotype, made of two genes: one gene aims at estimating the magnitude of the risk, the other the magnitude of the benefit. The difference between the values of the two genes defines the risk sensibility of the individual. The model is evolutionary because preys reproduce sexually and transfer the genotype to their offsprings. In this process, mutations can occur.
The analysis of the simulation results shows that the evolutionary process changes the distribution of genes in the population and can generate a risk-sensitive population. Moreover, it highlights that specific environmental factors influence the size and direction of the emergence of risk-sensitivity. A Bayesian network helps us to understand the relationships between environmental features and the emergence of risk sensitivity.
We find out that the riskiness of the environment has a non-monotonic influence on the risk sensitivity of the preys’ population. For low levels of riskiness, it is risk-averse, and the risk sensitivity grows with the riskiness. For high levels of riskiness, risk sensitivity decreases until the population becomes risk-seeking. Furthermore, we find observable relationships between the emergence of risk aversion in preys’ population and information process capacity, energy management, coordination ability and escaping strategy of preys’ individuals. Further extensions to this work could include explorations of risk aversion emergence in a coevolution prey-predator environment with more complex coordination strategies.