Monte Carlo Simulations : Performance and Insights

Danny PMP, PgMP

Community Champion

Senior Consultant Tokyo, Japan

Hi everyone, I recently conducted some Monte Carlo simulations, and I noticed they tend to take quite a bit of time to run, especially as the number of iterations increases. I'm curious has anyone here used Monte Carlo simulations in their work or projects? Did you face similar performance issues, and how did you handle them? Any tips or tools to speed things up? Would love to hear your experiences! Thank you.

Posted: Jul 20, 2025 9:10 PM (Updated by author: Jul 20, 2025 9:10 PM)

Sort By:

Sergio Luis Conte Helping to create solutions for everyone| Worldwide based Organizations Buenos Aires, Argentina

The key point is first of all understand how Monte Carlo simulations is implemented inside the tool you are using and then to understand the basement of scheduling method to use Monte Carlos simulations. I mean, it is not just to click a button and run a simulation to see what the results say about your schedule.

Posted: Jul 21, 2025 5:38 PM

Kiron Bondale Retired | Mentor| Retired Welland, Ontario, Canada

Danny -

I've used them a few times in the past for quantitative risk analysis and for forecasting likelihoods on completing a set of work items in a given amount of time. As computing power improved, the speed of simulations improved correspondingly although many of the popular tools (e.g. @Risk) added features which introduced some inefficiencies. How many iterations are you running as a few hundred are usually enough for most purposes?

Kiron

...

1 reply by Danny PMP, PgMP

Jul 28, 2025 4:22 AM

Danny PMP, PgMP

...

Dear Kiron Bondale , I'm running Monte Carlo simulations around 50K times. Anyway, thanks for your advice, it’s indeed very helpful.

Keith Novak & Sergio Luis Conte , appreciate for sharing as well. I finally managed to complete the simulations without any major issues. It just took a very long time. =)

Posted: Jul 22, 2025 7:21 AM

Keith Novak Tukwila, Wa, United States

Hi Danny,
There are a number of factors involved that will determine your processing time. As MC simulations start to include more variables and iterations, they can require huge amounts of memory being accessed and overwritten at each iteration. If you are trying to run them locally on a PC, the speed will vary greatly between a typical business PC vs. high end. Cloud based systems have vastly more memory, and some like AWS have MC algorithms included which simplify the process including creating distributions from your existing data.

Whenever I run processes that take a lot of time to execute, I typically run it like an experiment where I start with a few variables and lower iterations. That will give a feel for which variables dominate the output and how long each run will take. Otherwise you can get into situations where your process runs for many hours before it crashes the system. At university computing labs, I would often start them in the late evening and see if I had output in the morning, or if I crashed the mainframe again and then adjust from there to find a manageable size without losing a day for each overly optimistic attempt.
Keith

Posted: Jul 25, 2025 11:59 AM (Updated by author: Jul 25, 2025 12:00 PM)

Danny PMP, PgMP

Community Champion

Senior Consultant Tokyo, Japan

Jul 22, 2025 7:21 AM

Replying to Kiron Bondale

...

Posted: Jul 28, 2025 4:22 AM (Updated by author: Jul 28, 2025 4:28 AM)

Please login or join to reply

Monte Carlo Simulations : Performance and Insights

Sponsors

Vendor Events

Guessing is not a strategy: How to build decision velocity with AI and real-time data

Newsletters