I don’t think
- (lambda*norm(L,1) -lambda*(1/variance)*N)
winds up being part of the objective function.because it is not within the parentheses for maximize.
And
log_det(Precision_matrix)-trace(Covariance_matrix)*(Precision_matrix)
doesn’t even involve CVX variables or expressions, the way you’ve written the program. And Precision_matrix s dangling by itself (misplaced parentheses?) without trace or log_det being applied to it, which doesn’t make sense in an objective function. Is Precision_matrix supposed to be a CVX (optimization) variable? And if so, how does it relate to L? What is the relation (inverse?) between Covariance_matrix and Precision_matrix? You need to be very clear what is input data and what are the optimization variables and their relations. And for whatever you decide on that, it better be a convex optimization problem with respect to the optimization variables.