Your problem as stated is not convex, and not DCP-compliant as required by cvx.
Presuming Q is positive definite, then log_det(Q) is concave, and is therefore acceptable in cvx as an objective to be maximized. But it therefore is not acceptable to cvx as the left-hand side of a <= constraint.
Please review the cvx User’s Guide, and determine whether a) you actually wish to solve a non-convex problem, in which case cvx is not the right tool for it, or b) you are not formulating your problem correctly, and the problem you want to solve really is convex, in which case perhaps it can be formulated to be DCP-compliant and solved with cvx.
Edit: In response to the added question by the OP “Or are there any methods used to constraint the optimal value? Such as options in other toolbox.” If you want to constrain the optimal value and use cvx, then add only DCP-compliant convex constraints, otherwise, choose a tool other than cvx for solving your problem.