AutoRNet: Automatically Optimizing Heuristics for Robust Network Design via LLMs

Overview

AutoRNet is a framework that combines Large Language Models (LLMs) and Evolutionary Algorithms (EAs) to automatically generate robust heuristics for network optimization. It focuses on enhancing network robustness by leveraging domain-specific knowledge and adaptive strategies. This repository includes the implementation of AutoRNet, along with example datasets and tools for evaluating network robustness.

Key features:

•	Integration of LLMs with EAs for automated heuristic design.
•	Network Optimization Strategies (NOS) for domain-specific problem solving.
•	Adaptive Fitness Function (AFF) to balance convergence and diversity during optimization.

Getting Started

Prerequisites

•	Python 3.8+
•	Required libraries (install using pip)

Outputs

•	Optimized networks: Results are saved in the output folder or displayed in the visualization tool.
•	Robustness scores: Evaluations are logged during each run.

Key Modules

•	problem/: Defines network structures and constraints for optimization.
•	genetic.py: Implements the core EA logic, including variation and selection operations.
•	prompt.py: Generates problem-specific prompts for LLMs.
•	connector.py: Manages the interaction between LLMs and EAs.
•	viewer/: Provides visualization utilities for analyzing and presenting results.

Citation

If you use AutoRNet in your research, please cite our paper:

@article{HeYu2024, title={AutoRNet: Automatically Optimizing Heuristics for Robust Network Design via Large Language Models}, author={He Yu and Jing Liu}, journal={Preprint}, year={2024} }