I.6.2: Definition

Formalizing augmentation is straightforward. Suppose is a real valued function, defined for all , where Suppose there exists another real valued function defined on where such that We also suppose that minimizing over is hard while minimizing over is easy for all . And we suppose that minimizing over is also easy for all This last assumption is not too far-fatched, because we already know what the value at the minimum is.

I am not going to define hard and easy. What may be easy for you, may be hard for me.

Anyway, by augmenting the function we are in the block-relaxation situation again, and we can apply our general results on global convergence and linear convergence. The results can be adapted to the augmentation situation.

Note: augmentation duality.

then