AgenticProcessGNN

A multi-agent system leveraging Graph Neural Networks for process design, analysis, and optimization.

Overview

AgenticProcessGNN combines the power of Graph Neural Networks with a multi-agent framework to create an intelligent system for business process management. The system can analyze existing processes, design new processes, detect bottlenecks, and suggest optimizations.

Features

• Process Analysis: Analyze business processes using Graph Neural Networks
• Multi-Agent Framework: Collaborative agents for specialized tasks
• Vector Database: Store and retrieve process embeddings
• API Integration: REST API for system integration
• Process Optimization: Identify bottlenecks and suggest improvements
• Agent Memory: Long-term memory capabilities for agents to remember past interactions and domain knowledge

Architecture

The system consists of the following key components:

• GNN Model: Graph Neural Network for process analysis
• Agent Framework: Coordination mechanism for specialized AI agents
• Vector Store: Database for process embeddings
• API Layer: REST endpoints for system interaction
• Memory System: Persistent storage and retrieval of agent memories

Installation

Prerequisites

• Python 3.9+
• pip

Setup

1. Clone the repository:

   git clone https://github.com/yourusername/AgenticProcessGNN.git
   cd AgenticProcessGNN
   

2. Create a virtual environment:

   python -m venv venv
   source venv/bin/activate  # On Windows: venv\Scripts\activate
   

3. Install dependencies:

   pip install -r requirements.txt
   

Usage

Starting the server

python -m src.main

This will start the FastAPI server on http://localhost:8000.

API Documentation

Once the server is running, you can access the API documentation at:

• Swagger UI: http://localhost:8000/docs
• ReDoc: http://localhost:8000/redoc

Development

Project Structure

AgenticProcessGNN/
├── data/               # Data directory
│   ├── raw/            # Raw process data
│   ├── processed/      # Processed data
│   └── reference_processes/  # Reference process models
├── models/             # Saved model weights
├── src/                # Source code
│   ├── main.py         # Entry point
│   ├── agents/         # Agent framework
│   ├── process_engine/ # GNN model
│   ├── vector_db/      # Vector database
│   ├── api/            # API endpoints
│   └── memory/         # Agent memory system
├── tests/              # Test cases
├── docs/               # Documentation
├── requirements.txt    # Dependencies
└── README.md           # This file

Contributors

This project is developed by:

• Somesh Misra - ERP.ai Research
• Shashank Dixit - ERP.ai Research

License

This project is licensed under the MIT License - see the LICENSE file for details.

Citation

If you use this software in your research, please cite:

@software{AgenticProcessGNN,
  author = {Misra, Somesh and Dixit, Shashank},
  title = {AgenticProcessGNN: A Multi-Agent System with GNNs for Process Management},
  year = {2023},
  organization = {ERP.ai Research},
  url = {https://github.com/yourusername/AgenticProcessGNN}
}