Feature request: Monte Carlo optimisation with custom distributions

robertmartin8 / PyPortfolioOpt

Financial portfolio optimisation in python, including classical efficient frontier, Black-Litterman, Hierarchical Risk Parity

https://pyportfolioopt.readthedocs.io/

MIT License

4.35k stars 936 forks source link

Feature request: Monte Carlo optimisation with custom distributions #151

Open sabirjana opened 4 years ago

sabirjana commented 4 years ago

Hello Robert, I see "Monte Carlo optimization with custom distributions" in your road map. Definitely it will make this library highly flexible if there is an option to run Monte Carlo simulations for various kinds of optimizers offered by this library with added features to specify the kind of distribution. Please confirm my understanding of the road map you mentioned and if there is any update on progress towards this. Thanks Sabir Jana

robertmartin8 commented 4 years ago

Hi Sabir,

Appreciate the suggestion! I definitely need to do some deeper research into MC portfolio optimisation. The last time I read through, it seemed like people were just generating many random portfolios then selecting the best one out of that sample, which didn't seem like it was worth implementing.

If anyone has any input on what features a monte Carlo optimiser should offer, I'm all ears!

Best, Robert

naltang commented 3 years ago

Yes, the MC method may not look smart, but still worth implementing as a last resort when all other methods fail.

AdamTheDestroyer commented 3 years ago

Yes, it is THE possibility for Big Portfolios like I have (up to 7000 assets).

Greetings Adam

sidd1996 commented 2 years ago

@robertmartin8 @AdamTheDestroyer I will be happy to contribute to this open item,do let me know .