You may use anything from PWIQF2 and PWOQF2 as starting points. There are quite a few things in Excel and VBA that could form the basis for some things that are not formulae. (Bear in mind that I write for understandibility not for efficiency!)P

I have a C++ framework for 1 factor PDE/FDM models. At the moment it is based on the GOF patterns which is fine by me. A 2nd parallel version wil be based on interfaces (ICA) and interface contracts. Then we provide hooks that developers can plug in by implementing standardised interfaces. (Boost is essential imo).I see hooks for advanced PDE models, such as UVM, more types of PDE. Both 'engines' must be integrated into @rnorm's layers proposal.At some stage moving to ADI, Soviet Splitting and ADE will on the cards. (a bit of a short description, initial response) //What about Excel, C#, Mathematica intregration?

(Short description you say?)Apols Such but if you really want us to help, please do realise not all of us here are worldwide experts on PDEs.Can you please dumb down your specs a bit? I really would like to understand what you're saying, or else we'll need a Thesaurus. We come from various different backgrounds here and although this is a q-fin forum, it wouldn't hurt'compiling a dictionary of some sorts, so that we're talking about the same thing when we collaborate. An addenda to PW most frequently asked questions perhaps?Serp,

QuoteOriginally posted by: SierpinskyJanitor(Short description you say?)Apols Such but if you really want us to help, please do realise not all of us here are worldwide experts on PDEs.Can you please dumb down your specs a bit? I really would like to understand what you're saying, or else we'll need a Thesaurus. We come from various different backgrounds here and although this is a q-fin forum, it wouldn't hurt'compiling a dictionary of some sorts, so that we're talking about the same thing when we collaborate. An addenda to PW most frequently asked questions perhaps?Serp,Don't worry I will document it properly I(actually I already have). And I will take it in steps. And again, most has already been documented, really. It's not a high-risk item (but that's project management issue for me, thus a known).No problem. The project plan has yet to roll out.. PDE/FDM is in layer 2 (or 3) of rnorm's model? Quote@paul (Bear in mind that I write for understandibility not for efficiency!)We can use these models as input, which is no problem imo.

I will provide parallel Monte-Carlo framework and several fully implemented engines.

Interesting, could you just describe in more detail what exactly is meant by: parallel MC framework + engines.Let's take Glasserman for ref, how far does this contribution go?Cheers

QuoteOriginally posted by: renormI will provide parallel Monte-Carlo framework and several fully implemented engines.Also, is your code QMC ready? Unfortunately retrofitting in a framework thought for MC only causes complications with advanced QMC. I might help on this, time permitting...

QuoteOriginally posted by: outrunQuoteOriginally posted by: CuchulainnWhat about Excel, C#, Mathematica intregration?indeed, and Python.Hey, late at the party. I'd like to be in. We could use SWIG for this and for porting the lib to other scripting lingos (R and Ruby are also supported).

QuoteOriginally posted by: renormI will provide parallel Monte-Carlo framework and several fully implemented engines.I'm sure this is gonna be the best code I've ever seen . I think the MC framework will be very very important -- maybe we should discuss it extensively in a specific thread.