L-One Norm of Derivative Objective
Another way to add smoothness constraint is to add -norm of the derivative to the objective:
(4.82) |
Note that the norm is sensitive to all the derivatives, not just the largest.
We can formulate an LP problem by adding a vector of optimization parameters which bound derivatives:
(4.83) |
In matrix form,
(4.84) |
The objective function becomes
(4.85) |
See Fig.3.41 and Fig.3.42 for example results.
Next Section:
Summary
Previous Section:
L-Infinity Norm of Derivative Objective