First of all: I am an absolute beginner in terms of convex optimization and DCP.
I am using CVX to solve the following problem:
minimize: norm(A*x - b)
subject to: 0 <= x <= 1
Being a part of an iterative algorithm I try to introduce some kind of penalty Term to get the elements of x to converge to 0 and 1 respectively. The binarization term is meant to have an higher impact by increasing a weighting factor lambda with every iteration. So CVX shall solve the problem in every iteration of the algorithm with an increased value of lambda until x persists only of 0s and 1s.
I can think of various ways to mathematically formulate this but I was unable to translate this into DCP so far.
Can anyone of you help me out with this?
Thank you very much for all replies in advance