I have fabricated embedded optical waveguides in fused silica using femtosecond laser micromachining with a permutation of various laser parameters. When I characterize these fabricated waveguides using the butt-coupling of waveguide end with a single mode fiber at a certain wavelength, the waveguiding phenomenon occurs. However, for various laser fabrication parameters, the measured propagation losses varies. To find the most optimum combination of fabrication parameters, is there a computational route? Since the core diameter is an unknown and the refractive index change is generally within a range of 10^-3 - 10^-4, how do I computationally or experimentally optimize to find a window of parameters which would give me low losses without fabricating and characterizing all of the permutations of waveguides?