In this paper we study several means of compensating for thermal lensing wh
ich, otherwise, should be a source of concern for future upgrades of interf
erometric detectors of gravitational waves. The methods we develop are base
d on the principle of heating the cold parts of the mirrors. Ve find that t
hermal compensation can help a lot but can not do miracles. It seems finall
y that the best strategy for future upgrades ("advanced configurations") is
may be to use thermal compensation together with another substrate materia
ls than silica, for example sapphire.