Hello, I was wondering if in 1d) D and C were still supposed to be independent, or if we should treat the maximization problem in itself, with C having no link with the scope of the problem (likewise is C still nonnegative a.s.?).
Thank you
Hi !
C should fulfill the scope of the problem (i.e., C independent of D, C non-negative a.s., expectation of C less than L-1/lambda), so it gives a valid lower bound on the performance of the scheme.
Best,
Reka
