Multilevel Monte Carlo Support in QuantLib

Summary

The document asks whether QuantLib implements multilevel Monte Carlo and whether adding the method would fit the library's design. The answer states that the library has no multilevel Monte Carlo implementation at the time discussed. It suggests that one could be structured as a simulation class modeled on the library's existing Monte Carlo simulation framework, assembled from available components.

This is a concise library implementation note, not an explanation of the multilevel Monte Carlo algorithm or its numerical advantages. It provides no benchmark, code, implementation plan, or discussion of the method's suitability for particular pricing problems. The suggestion that existing building blocks could be combined indicates a possible architectural path, but feasibility and performance would need to be assessed in an actual implementation.

Key ideas

• The answer reports that QuantLib does not include multilevel Monte Carlo.
• It suggests a dedicated simulation class as a possible implementation shape.
• The proposed class could combine existing library components.
• The document does not describe the algorithm, implementation details, or performance evidence.

Tags

# Is there Multilevel Monte Carlo in QuantLib?


# Is there Multilevel Monte Carlo in QuantLib?












Is the Multilevel Monte Carlo method implemented in QuantLib? If not, would it make sense to implement it? Is it doable taking into account the structure of the library?

## Answer by Luigi Ballabio (score 3)

https://quant.stackexchange.com/a/78307

There's no multilevel MC in the library. I guess it could be implemented as a class similar to `McSimulation` (see this link) which puts together a number of available building blocks.

Shown in full with attribution under the source's licence. Licence: CC BY-SA 4.0 (Stack Exchange)

This summary was written by Stratmill's research agent from the original; it is not a copy of the source.