How to represent this set of positive semidefinite matrices into a cvx variable?

[hancheng_zhu]

I have a set of matrices {V_{k,l}}, where k \in \left\{ {1,2, \cdots ,K} \right\} and l \in \left\{ {1,2, \cdots ,L} \right\}. Each {V_{k,l}} is a N*N dimensional matrix. I want to represent it into a cvx variable. I know when only k exists, i can declare it as

variable V(N,N,K) complex semidefinite

But it has two indices k and l. I know i can declare it as

variable V(N,N,KL) complex semidefinite

where KL=K*L, but it is a little inconvenient when i want to extract a V_{k,l} from this variable set V because i need to recover the third dimension of cvx variable V back to the two dimensions k and l. Is there any declaration like

variable V(N,N,K,L) complex semidefinite

maybe exist?

[Mark_L_Stone]

it should work for any number of dimensions.

How to change the number of variables based on the iteration number?

Why not use multidimensional arrays? For instance: variable Q_1(n,n,k) hermitian; variable Q_2(n-1,n-1,k) hermitian; Then you can access an individual ‘slice’ with colon notation; e.g., Q_1(:,:,k). From the documentation: Matrix-specific keywords can be applied to n-dimensional arrays as well: each 2-dimensional “slice” of the array is given the stated structure. So for instance, the declaration variable R(10,10,8) hermitian semidefinite constructs 8 10×10 complex Hermitian PSD matr…