I’d like to minimized the following problem:
minimize( norm( R*x, a ) ),
but rather than having CVX calculate R*x directly, I’d like to supply a function handle that does it with a more efficient algorithm:
minimize( norm( fhandle(x), a ) )
Is this possible? I am getting complaints about unsupported argument type ‘cvx’ and have not been able to figure out how to cast x back to a double. I saw the section on using ‘expressions’ in the user guide but can’t see how that would help in this situation.
Edit: What if R is an object with the ‘*’ operator overloaded?