Skip to content

Verifying closed-form solutions through the use of simulation and the law of large numbers.

Notifications You must be signed in to change notification settings

Booleans/math-through-simulation

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

63 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Math Through Simulation

Vampires

Thanks to the power of modern computers and the law of large numbers, certain math problems can be solved through simulation. Take this data science interview question as an example:

Every night between 7 PM and midnight, two computing jobs from two different sources are randomly started with each one lasting an hour. Unfortunately, when the jobs simultaneously run, they cause a failure in some of your company's other nightly jobs, resulting in downtime that costs $1,000. The CEO needs you to tell her the annual cost of this issue.

With the help of (Python) code you can simply simulate these nightly jobs tens of millions of times and see how often they tend to overlap.

import numpy as np
# There are five hours in this interval and each job lasts one hour, or 20% of the total interval.
# Let's use np.random.random to generate a number between 0 and 1, representing the position in the interval that the jobs start.
# If the difference between start times is <= 20% of the interval then there's an overlap. 
process_1 = np.random.random(size=10**7)
process_2 = np.random.random(size=10**7)
overlap_percentage = np.mean(np.abs(process_1 - process_2) <= 0.20)
annual_cost = overlap_percentage * 365 * 1000

NumPy Logo

About

Verifying closed-form solutions through the use of simulation and the law of large numbers.

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages