Dynamic World Story using LLM Agent-Based Simulation

✨ Crafting Emergent Worlds with AI Agents

This project explores a novel multi-agent simulation system that leverages Large Language Models (LLMs) to generate dynamic, emergent narratives. A unique "Director" agent subtly guides the evolving story by altering the environment and situations of agents, fostering a rich flow of perception, reaction, and plot developments without direct intervention.

Showcases deep expertise in LLMs, Agent-Based Modeling, and advanced AI applications.

Key Technical Achievements & Features

Core Capabilities:

• LLM-Powered Agent Architecture: Designed and implemented intelligent agents whose internal states, decision-making, and interactions are dynamically driven by integrated Large Language Models, enabling complex and adaptable behaviors.
• Agent-Based Simulation Framework: Built a robust simulation environment capable of managing multiple concurrent agents, their shared world state, and the progression of time, providing the core for complex interactive scenarios.
• Emergent Story Generation: Explored and implemented principles of emergent storytelling, observing how complex narratives unfold organically from simple agent rules and LLM-driven interactions, rather than static plots.
• Advanced Prompt Engineering: Utilized sophisticated prompt engineering techniques to define agent personalities, manage context, and elicit coherent and creative responses from the LLMs, crucial for consistent narrative flow.

Project Structure

• src_GM/ - Main source code directory containing the simulation framework
  • main.py - Entry point for running simulations
  • agent/ - Agent implementation with memory and planning systems
  • director.py - Director agent for guiding story development
  • world.py - World state management
  • config.py - Configuration settings
• initial configs/ - Pre-configured story scenarios
• All Stories/ - Generated story outputs and examples
• objectives.txt - Project goals and development notes

Getting Started

1. Configure your LLM API keys in src_GM/config.py
2. Choose a story configuration from initial configs/ or create your own
3. Run the simulation: python src_GM/main.py
4. Watch as AI agents interact and create emergent narratives!

About This Project

This project represents a significant contribution to Artificial Intelligence, Large Language Models, Agent-Based Modeling, and Computational Creativity, demonstrating both theoretical understanding and practical implementation skills in creating systems where complex stories emerge from simple agent interactions.