Not equal RHS and LHS with norm function

Hello. I have a expression in CVX, after applying the norm function on it, the answer of both sides of the equality is not the same. You answered in previous posts that we should use the variable of CVX and then computing the expression.

But is there a problem with CVX in this matter as well? alti is a CVX variable.

Thank you very much for your guidance.

QWEERT is an expression. Therefore, to find it’s optimal value, you must compute it after cvx_end, using CVX variable values, which in this case according to you is alti Therefore, 18.2079 is perhaps the correct value.

However, you have not clearly shown what your program is, nor the solver or CVX output, nor the order of any supplemental calculations. If the “formula” for QWEERT is defined between cvx_begin and cvx_end;, it would be an expression, and what I said in the first paragraph should be true.

The value of 5e13 is extremely large, and the fact that it might be produced in any circumstances makes me wonder about the problem scaling, to include the numerical scaling of the input data, which might cause various difficulties.