[Quantlib-users] Local vol and Andreasen Local vol implementation

[Dan V. <dan...@gm...>]

Hello,

We try to price a MC product on equity with local volatility. In this
regard, I first have a black variance surface with holes from implying vols
in the market.

Of course simply activating local vol on that surface is not sufficient as
many arbitrages are found, therefore this is not a robust enough solution.
FYI I am providing a very granular surface because I need precision at
every time step to the downside.

We tried using the AndreasenHugeLocalVolAdapter as it is supposed to
provide an arb free workable local vol surface. But several things:
1. If used with a basic vanilla pricer (for testing purposes) it still
finds arbitrages and can't price
2. It is unclear whether we still need to build a Vol Term Structure object
and if we do, is there such a thing as a localvoltermstructure object ? if
there is, it would be great to know the syntax as we couldn't find anything
on it on the web.
3. Lastly, if we manage to fix 1 & 2, how do we go about building a path on
the back of that term structure. I couldn't find any localvol processes.
Should we build a BlackMerton process except feeding it a localvol term
structure ? Will it understand it ?

Basically any sample code to generate local vol paths for MC pricing would
be greatly appreciated.

Thanks,
Dan

PS: we are using Python