Proteus Actuarial Library

An actuarial stochastic modeling library in python.

Note This library is still in beta!

Introduction

The Proteus Actuarial Library (PAL) is a fast, lightweight framework for building simulation-based actuarial and financial models. It handles complex statistical dependencies using copulas while providing simple, intuitive syntax.

Key Features:

• Built on NumPy/SciPy for performance
• Optional GPU acceleration with CuPy
• Automatic dependency tracking between variables
• Comprehensive statistical distributions
• Clean, Pythonic API

Quick Start

from pal.copulas import GumbelCopula
from pal.distributions import Gamma, LogNormal

# Create stochastic variables
losses = Gamma(alpha=2.5, theta=2).generate()
expenses = LogNormal(mu=1, sigma=0.5).generate()

# Apply statistical dependencies
GumbelCopula(theta=1.2).apply([losses, expenses])

# Variables are now correlated
total = losses + expenses

Installation

# Basic installation
pip install proteusllp-actuarial-library

# With GPU support
pip install proteusllp-actuarial-library[gpu]

Documentation

Read the full documentation on Read the Docs

• Usage Guide - Comprehensive examples and API documentation
• Development Guide - Setting up the development environment and running tests
• Examples - Example scripts showing how to use the library

AI and coding assistants

PAL provides dedicated entry points for AI tools as well as human-readable documentation:

• PAL quick reference for coding assistants maps common modelling intentions to the public Python API.
• AGENTS.md contains canonical repository instructions for coding agents, with more-specific guidance under src/pal/, tests/ and docs/.
• The documentation build publishes llms.txt and llms-full.txt at the documentation root for model-friendly retrieval.

AI-generated PAL code should prefer documented public APIs and inspect signatures/docstrings rather than guessing parameterisations or relying on internal helpers.

Project Status

PAL is currently in early release preview (beta). There are a limited number of supported distributions and reinsurance contracts. We are working on:

• Adding more distributions and loss generation types
• Making it easier to work with multi-dimensional variables
• Adding support for Catastrophe loss generation
• Adding support for more reinsurance contract types (Surplus, Stop Loss etc)
• Stratified sampling and Quasi-Monte Carlo methods
• Reporting dashboards

Issues

Please log issues on our github page.

Contributing

You are welcome to contribute pull requests. Please see the Contributer License Agreement

📚 Development Guide - Get started with development setup and testing