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?
Yes, quantum computers can efficiently simulate quantum systems, it is really very much like the situation with NP-complete or NP-hard problems. [1]
As far as we know and if we did not miss a loophole, the violation of Bell's inequality force us to give up locality, realism or freedom. So you can keep locality if you throw realism or freedom over board. But this will (probably) not help winning the money in practice.
Also, are you sure that the Copenhagen interpretation is correct? Aren't 'local hidden variable theory' and 'consistent histories theory' also possible?