I had this need but never felt compelled enough to tackle the problem.. There are two different issues here.
The first one is how to deal with jumps. The framework is ready to deal with jumps as exogenously superimposed over a smooth function, as usually needed by turn-of-year effects and similar jumps. That is the way to go in my opinion for local jumps.
The second one is when you might prefer something like piece-wise-flat for the first two years then cubic. In this case some work has been done on mixed interpolation, which might be useful for instantiating just one curve with a mixed interpolation. In that case you have the complexity of defining the interpolation traits.
It might be probably better to define instead a new term-structure which does the blending of two different term structures. It is less efficient in some way, but probably more robust.
------------------------------------------------------------------------------
Put Bad Developers to Shame
Dominate Development with Jenkins Continuous Integration
Continuously Automate Build, Test & Deployment
Start a new project now. Try Jenkins in the cloud.
http://p.sf.net/sfu/13600_Cloudbees_______________________________________________
QuantLib-users mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/quantlib-users
Locked
