Adding optimization example

[nestordemeure]

Hello,

I was searching for a Rust implementation of LBFGS for a personnal project and your crate seem to perfectly fit the bill.

However, while you document how to update the Hessian and correct a gradient, I am unsure whether I should update then correct or correct then update or avoid correcting when I am updating...

Adding an example with a very small optimization loop (minimizing a function with known gradient for a fixed number of iterations) would help potential users a lot.

[korken89]

Thank you for the idea!
I think this can be arranged, I will give it a look.