Evolution of Fear and Social Rewards in Prey-Predator Relationship

TL;DR

The study addresses how predation pressure and social rewards shape the evolution of fear in a prey–predator system by using a distributed evolutionary RL framework where prey and predator reward functions co-evolve under energy-based birth–death dynamics. The main approach combines agent-based simulations with co-inherited reward weights, PPO-based learning, and environmental manipulations such as predator mouth size and resource regeneration rate $\Delta n$ to study fear and social reward emergence. Key contributions include the demonstration that fear rewards frequently evolve in prey, that social rewards can co-evolve with fear under certain conditions, and that predator lethality and movement intensify fear evolution, while static pitfalls do not reliably induce fear. The work offers insights into the evolution of complex emotions within social contexts and illustrates how multi-agent evolutionary simulations can reveal strategies beyond traditional optimality frameworks, with implications for understanding natural behaviors and informing AI systems.

Abstract

Fear is a critical brain function that enables us to learn to avoid danger via reinforcement learning (RL). While many researchers have argued that fear has evolved to escape predators, how varying predatory pressures have shaped fear and other rewards, including positive social rewards for collective grouping, remains an open question. In this study, we investigate the relationship between predatory pressure and fear using an evolutionary simulation of RL agents with evolving rewards. In our simulation, prey and predator RL agents co-evolve their reward functions, including visual rewards for observing prey and predators. While fear-like negative visual rewards for predators often evolved in prey, we also observed cases in which positive rewards for both predators and prey evolved, the latter serving as a social reward for collective grouping. A comparison between different environmental conditions revealed that stronger predator hunting capability promoted stronger fear reward, while less food supply promoted more negative social reward. Moreover, fear did not evolve in response to static pitfalls with non-lethal damage, suggesting that actively hunting predators played an important role in its evolution. These results highlight the special role of predators in the diverse evolution of fear and social rewards.

Figures (22)

• Figure 1: Model of prey and predators used in our simulation. Blue circles represent prey, red circles represent predators, and green circles represent food items that only prey can eat. Gray lines extending from the prey are proximity sensors.
• Figure 2: (a): Environment with prey and predators. (b) Environment with prey static pitfalls (black triangles).
• Figure 3: Mouth ranges of prey and predators. M is used as the default value in most experiments.
• Figure 4: Schematic figure shows the learning procedure of prey and predators.
• Figure 5: Simulation procedure.