File Exchange

image thumbnail

Pseudorandom number generation for engineering estimates

version 1.1 (4.94 KB) by Sky Sartorius
Tools for generating pseudorandom numbers, primarily geared toward making engineering estimates.

2 Downloads

Updated 13 Feb 2019

GitHub view license on GitHub

A tool for random number generation on a distribution, such as a triangular distribution or PERT distribution, is very convenient for making estimates for physical values in the real world. Scientists and engineers often make estimates or assumptions, but they are also able to bracket the estimate based on experience or physics. For example, in estimating the mass of a small object, a normal distribution centered at my best guess could result in negative mass, which is known to be impossible. A triangular distribution allows quick and simple characterization of values as "probably M, but definitely not less than A or more than B."

The formulations here have some features that make it easy to integrate this capturing of uncertainty into scripts.

Example: Based purely on guesstimates that include a best guess and intuitive upper/lower limits, how much do a dime, nickel, and quarter weigh together?

% First, create a function based on randt to conveniently define and generate "uncertain variables."
uvar = @(x) randt(x,[1e5,1]);

dime = uvar([1 1.5 3.5]); %Pretty light, but not less than a gram..."
nickel = uvar([3 5 6]);
quarter = uvar([5 8 10]);

total = dime + nickel + quarter;

histogram(total,'Normalization','pdf');

Cite As

Sky Sartorius (2019). Pseudorandom number generation for engineering estimates (https://www.github.com/sky-s/randx), GitHub. Retrieved .

Comments and Ratings (1)

Updates

1.1

Added PERT distribution. Moved to GitHub repo.

MATLAB Release Compatibility
Created with R2016a
Compatible with any release
Platform Compatibility
Windows macOS Linux