I just noticed that Professor Mark Joshi passed away last month, and I have really benefited from his book C++ Design Patterns and Derivatives Pricing when I was dong my master’s degree.
I was wondering if it is possible for Julia to gradually replace C++ in the area of financial numerical computation? I noticed that QuantLIb.jl is under working, which is a good sign, and I am looking forward to seeing more developments in this area.
Yes. I plan to keep on releasing a bunch of numerical SDE stuff only in Julia, and we are already very far ahead in that domain. It’s really easy to write PDE solvers in Julia, so we should soon get a plethora of options here to then plug into. I don’t think the old school QuantLIb style is good at all, because they offer things like inefficient PDE solvers since it’s just one piece of their offer, instead of grabbing a more robust PDE solver. I think a Julia-based financial modeling package could tie together disparate libraries and put a financial spin on it, so it could just use already optimized code but make it easier for quants to use it. That would not only be much easier to develop, but it should be vastly more efficient than QuantLib.
Are there any financial packages that use the efficient SDE solver you mentioned?
Don’t know about financial. I know systems biology and particle physics researchers have been making use of it. But since it’s just an SDE solver, it’s not domain-specific at all.
I made some very basic sample problems from finance to test on:
http://docs.juliadiffeq.org/latest/models/financial.html
but I don’t think I myself will really build DiffEqFinancial.jl to be a huge modeling package since it’s not my domain. I will however keep on writing ODE/SDE/PDE/etc. solvers and help anyone who wants to build such a modeling package.