Suppose we have a complex refractive index at a given frequency from experimental data (the imaginary part being responsible of loss/gain mechanisms).
If we are to use time domain electromagnetic solver (FDTD or DGTD) one way to incorporate gain/loss is through the use of the polarization equation, which expressed in the frequency domain gives a linear relation between P and E. The linear coefficient is the called the susceptibility and has the form of a second order filter with a dampening coefficient gamma.
The question is then how to choose this gamma coefficient so that we gain the same gain/loss at the resonant frequency as what we would get by using the refractive index?