Quadratic Programming & Multi-Dimensional Newton Iteration

Posted by tallent_e on May 30, 2015; 11:39am
URL: http://quantlib.414.s1.nabble.com/Quadratic-Programming-Multi-Dimensional-Newton-Iteration-tp16604.html

Hi Vivian

Luigi Ballabio commented (in reply to a question from Mark Joshi) that QP factories would probably be on the todo list some day.

It was a couple years ago, now.

Yet there is still no such tools available in QL, to my knowledge.

 

Best

Edouard.

 

+++

 

Asunto	[Quantlib-users] Quadratic Programming & Multi-Dimensional Newton Iteration
Remitente	[hidden email]
Destinatario	[hidden email]
Fecha	Vie 09:30

Hi, all,

An optimization problem can be formulated as follows. 
Take 2-Dimensional as an example(variables: x1, x2; constants: a~i).
min a*pow(x1,2)+b*x1*x2+c*pow(x2,2)
s.t.         d*x1+e*x2+f=0
              g*x1+h*x2+i>=0
              x1>=0
              x2>=0
We could use quadratic programming or multi-dimensional Newton iteration to
solve it.

After going through QuantLib to search for quadratic programming or
multi-dimensional Newton iteration, I found OptimizationMethod with
CostFunction and Constraint(boundary constraints of the variables but not
linear constraints on these variables) and Newton 1-D solver(not
multi-dimensional). So they are not suitable for the problem.

Did I misunderstand these classes, like OptimizationMethod, and Newton?
Or I have missed the functions related to quadratic programming?
Is there any other ideas about how to solve the above problem?


Thanks in advance.


Regards,

       Vivian

++++

------------------------------------------------------------------------------

_______________________________________________
QuantLib-users mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/quantlib-users