Rename to avoid overwriting in loop

[SJaffa]

Closes #378

I'm not sure initial_res is the most sensible name here but @krishnakumarg1984 it does build and pass all the tests.
@mmcleod89 you read the papers most recently, do you have any idea what this res variable is and what a sensible name might be?

[mmcleod89]

I am struggling to figure out what precisely res is... usually it means result or residual, and it doesn't appear to be a residual. We use Psi.adjoint to convert from the image domain to the wavelet domain, apply some threshold and scaling, and then convert back to the image domain and subtract it from x so it appears to be some kind of intermediate value that is used in the udate, but I can't quite figure out how it relates to the papers right now - it doesn't help that the use of nu, gamma, lambda, x, z etc. is very inconsistent!

Another question that I think this change raises is why are we making these const variables in the first place? What is the point of keeping initial_res given that it's never used again? This is an iterative method, and the initial value is just declared outside the loop. Wouldn't it make more sense to declare it non-const and just update it every iteration and re-use the memory rather than creating and destroying the vector every iteration? (Granted it might be optimised out, but I still don't see why we wouldn't just use a non const variable to begin with.)