/psyche

AI Agents Society Simulation

Primary LanguagePythonMIT LicenseMIT

Psyche: An Agent Society Simulation

Psyche is an experimental platform for exploring agent societies and emergent behavior. The simulation models interactions between agents with diverse personalities, needs, and goals, creating a dynamic environment where agents evolve through cooperation, competition, and adaptation.

Core Features

  • Agent Personalities: Each agent has distinct traits and decision-making processes.
  • Resource Management: Basic resources like food and gold drive survival and interactions.
  • Dynamic World Events: A central “World Agent” introduces random events to shape the environment and test adaptability.
  • Emergent Behavior: Agents act based on their personalities and situations, leading to unexpected outcomes.

First Experiment: The Village

The initial experiment, The Village, simulates a small community where agents must balance resource management, trade, and survival. It serves as a testbed for observing how simple rules and interactions can lead to complex societal dynamics.

Goals

  • Understand how agents adapt to limited resources and environmental changes.
  • Explore the formation of relationships, alliances, and conflicts.
  • Test the emergence of group behaviors from individual decision-making.

Future Directions

  • Expanding agent roles and personalities.
  • Introducing long-term evolution and learning.
  • Simulating larger, more complex societies with diverse goals.

Psyche is a foundation for studying how simple rules and interactions create complex systems, offering insights into both artificial and real-world social dynamics.