I recently saw an paper that performed successive convex approximation operations on a non-convex problem, but the proof of its convergence is very simple. (The specific situation is shown in the figure)
Please help me to see if this algorithm can give a strict mathematical proof of convergence?