from website:
The suite of fast incremental algorithms for machine learning (sofia-ml) can be used for training models for classification, regression, ranking, or combined regression and ranking. Several different techniques are available. This release is intended to aid researchers and practitioners who require fast methods for classification and ranking on large, sparse data sets.

Binomial pricing Tree

It can be used to price almost any exotic structure.

Complex structures like Worst of Down and In Put (WODIP), and several other combinations.

****Monte Carlo Pricing Framework*******Heston Model*****
Correlation Matrix:
1 0.999999 0.999999
0.999999 1 0.999999
0.999999 0.999999 1

Number of Sims:5000
Barrier: 120Strike: 100

*****Price List*****
*************************
Heston Model Parameters:
Kappa:1
Theta:0.04
V0: 0.04
eps: 0.03
Vol Correlations

*****************************
Price for Call on Basket 9.20366
Price for Put on Basket 5.84145
Price for Worst of Call 5.83126
Price for Worst of Put 11.0355
Price for Best of Call 10.6381
Price for Best of Put 5.95819
Price for WODIC 10.6766
Price for WOUOC 1.62048
Price for WODOC 0
Price for WOUIC 9.34861
Price for BODIC 10.6944
Price for BOUOC 1.64875
Price for BODOC 0
Price for BOUIC 8.87429
Price for WODIP 5.78065
Price for WOUOP 5.6988
Price for WODOP 0
Price for WOUIP 0.152209
Price for BODIP 5.7921
Price for BOUOP 5.65533
Price for BODOP 0
Price for BOUIP 0.155745
Price for:Up Out Call 1.65727
Price for:Down Out Call 9.41743
Price for:Down In Call 0
Price for:Up In Call 10.9601
Price for:Up Out Put 5.68174
Price for:Down Out Put 0.163071
Price for:Down In Put 0
Price for:Up In Put 5.78814

It prices almost all kinds of Barriers, Baskets and Complicated Basket structures.

The design in based on just adding an extra hook for new exotic structure. I have added hooks for pricing structures like Worst of Down and In Put. Simulations are done using Halton numbers and convergence is extremely fast. It takes less than 30 seconds to conduct 30,000 simulations.
Cholesky decomposition is used for generating correlated random numbers.
Volatility process / Stock correlations are also given by the user. They are different for each component of the basket.

I have used Heston Model for pricing. Correlation matrix is input by the user.

For any questions do mail me

animesh.saxena@gmail.com

I had coded this in XCode (Mac development GUI), but it can be compiled on any platform. You just need to assemble the files in a project! Enjoy!
Might be added to QuantLib library later (if developers @ QuantLib agree!), coz QuantLib still lacks a generic pricing framework.

PS: I have left out discounting, coz discounted price for deterministic rate will be genrally e^-rT * Expected value of Payoff
Will add Heath Jarrow Morton model later for adding interest rate calibrations.

ASA technique was developed by Lester Ingber to facilitate calculation of global minima for a nonlinear function. For more detailed information about the algorithm and it's usage, please refer to the author's Author's website

One important application is calibration of heston model, because the parameters are correlated and other numerical optimization methods tend to provide solution which is more sensitive to selection of initial parameter values.
This project implements calibration of Heston models using Adaptive simulated annealing.
To run this project,do the following
1.unzip the attached download file and extract the testheston project folder
2.download project files from file http://downloads.sourceforge.net/asa-caltech/ASA-26.33.zip?modtime=1230643798&big_mirror=0. Place the file asa.h and asa.c in the testheston->testheston folder.
3.Open project
testheston.sln in Visual C++ 2005 express
4.Do a build all.
5.Select run
6.The results of calibration are in file asa_out.

To play around with ASA optimization, make changes to initial parameters and other settings in asa_opt file

The function GetHestonCallprice returns result of calculating call price of a plain vanilla option according to heston's analyticla formula using numerical inegration