Suppose the convex optimization problem is

Variable x (5).

Min ∑ x (i) * c1/c2.

sum= ∑ (rel_entr (x (i), x (i) + 1) + rel_entr (x (i) + 1，x (i)）;

Subeject to (1): c3 < = x (i) < = c4;

(2): sum < = c5;

Why does this happen when the “sum” calculated within cvx_begin and cvx_end is inconsistent with the sum calculated by x (i) at the end of the convex optimization process?