[Quantlib-dev] 答复: QuantLib Solver

[Cheng L. <scr...@gm...>]

Hi Nigel,

 

Of cause it could. If you look at the implementation of the solver, the solve method’s first parameter is marked as template. You only need to make sure the VolatilityFunc accept a double and return a double.

 

Regards,

Cheng

 

发件人: Nigel Sperinck [mailto:nig...@gm...] 
发送时间: 2015年5月12日 6:20
收件人: qua...@li...
主题: [Quantlib-dev] QuantLib Solver

 

QuantLib Developers,

 

Reference

http://quantlib.org/slides/dima-ql-intro-2.pdf, page 23

 

In Dimitri Reiswich's QuantLib Solver example, he sets up a boost function and binds this to his function impliedVolProblem() which returns the calculation of the Black Scholes Option Price - price.

 

Are the following two steps required, or is there a way to refactor and simplify the code and use a function pointer directly in the QuantLib Solver.solve method without having to use the boost::function and boost::bind?

 

// Step 1 - Setup a boost function

boost::function < Real(Volatility) > VolatilityFunc ;

 

// Step 2 - Bind the boost function to all market parameters, keep volatility as variant

VolatilityFunc = boost::bind(&impliedVolatilityProblem, spot, strikeprice, rd, rf, _1, tau, phi, price);

 

Real res1 = bisectionSolver.solve(VolatilityFunc, accuracy, guess, min, max);

 

Very best regards

Nigel Sperinck