Well, I didn't specify what type of computer must be used. :-) It should be possible with a quantum computer, yea? Then it merely becomes rewriting one NP-hard problem in terms of another one (one being related to the refraction of light in glass and the other being whatever basis the quantum computer is built in terms of).

Also, are you sure that the Copenhagen interpretation is correct? Aren't 'local hidden variable theory' and 'consistent histories theory' also possible?