QMClib

Monte Carlo & Quasi-Monte Carlo Library

QMClib is an open-source Python library for numerical integration using Monte Carlo and quasi-Monte Carlo methods. Developed during my research internship at the DAO Team (Data, Learning and Optimization) of LJK Laboratory.

Features

  • Sampling Methods: Uniform, Latin Hypercube Sampling (LHS), Halton, Sobol sequences, and lattice-based methods
  • Statistical Analysis: Convergence evaluation and performance benchmarking
  • Modular Design: Easy to extend with new sampling methods and integration algorithms
  • Validation: Extensively tested and benchmarked against established libraries

Technical Details

The library is built with:

  • Python with NumPy and SciPy for numerical computations
  • Matplotlib for visualization
  • Pyperf for performance benchmarking

Applications

QMClib was applied to a machine learning risk estimation problem, exploring the potential of quasi-Monte Carlo methods for high-dimensional integration tasks.

This project was supervised by Quoc-Tung Le at LJK Laboratory, Grenoble, France (May 2026 - July 2026).