Thanks for the various replies. I think I will use a benchtop power supply with a current readout to find out how much current is required. The PWM-op amp-mosfet idea is a nice one and I had not thought of a battery charger as a model to follow. Although the response time and sensitivity of a PWM approach is not going to be fantastic, this shouldn't be an issue because once set, it is not likely to be changed. I have a Pockels cell to control the power and that will be controlled with a DAQ device with a much faster update rate and sensitivity.
Thanks,
Nick