Skip to content

mcerp

MCERP

Real-time latin-hypercube-sampling-based Monte Carlo error propagation for Python.

mcerp is a stochastic calculator for uncertainty analysis. It represents inputs as probability distributions, samples them with Latin hypercube sampling, and carries those samples through ordinary Python calculations. The result is another uncertain value whose mean, variance, skewness, and kurtosis can be inspected directly.

If you are familiar with Excel-based risk analysis programs like @Risk, Crystal Ball, ModelRisk, etc., this package will work wonders for you (and probably even be faster!) and give you more modelling flexibility with the powerful Python language. This package also doesn't cost a penny, compared to those commercial packages which cost thousands of dollars for a single-seat license. Feel free to copy and redistribute this package as much as you desire!

Use it when a model is easier to express as code than as a closed-form error formula:

  • manufacturing tolerance stack-ups
  • engineering calculations with uncertain measurements
  • reliability and risk calculations
  • probability estimates from simulated outputs
  • quick exploration of how input uncertainty changes a result

Why mcerp?

mcerp lets you write calculations almost as if every uncertain input were an ordinary number:

from mcerp import Exp, N

x1 = N(24, 1)
x2 = N(37, 4)
x3 = Exp(2)

Z = (x1 * x2**2) / (15 * (1.5 + x3))
Z.describe()

The output summarizes the distribution produced by the calculation:

MCERP Uncertain Value:
 > Mean...................  1161.35518507
 > Variance...............  116688.945979
 > Skewness Coefficient...  0.353867228823
 > Kurtosis Coefficient...  3.00238273799

Main Features

  1. Transparent uncertainty propagation through normal arithmetic.
  2. Constructors for common continuous and discrete probability distributions.
  3. Support for many mathematical functions through mcerp.umath.
  4. Probability calculations with comparison operators such as <, >=, and ==.
  5. Correlation enforcement and correlation matrix utilities.
  6. Optional distribution plotting with Matplotlib.
  7. Direct compatibility with scipy.stats distributions through uv(...).

Documentation

  • Theory explains Monte Carlo propagation, Latin hypercube sampling, moments, and correlations.
  • Installation shows package installation options.
  • Quickstart walks through the core workflow.
  • API Reference lists the public constructors and helpers.
  • References collects the source material and related projects.