As long as you can introduce in your code observable of end-to-end vector and trace its autocorrelation function, this will give you exactly what dielectric spectroscopy does.

One method we and others have used is to apply a 1939 theory from Kirkwood relating a time/ensemble average of dipole moments within a given radius to the dielectric constant. We used it to calculate local dielectric constants in different regions surrounding a DNA molecule, using data from an AMBER MD simulation of DNA in TIP3P water. I believe this RG link will take you to the paper: https://www.researchgate.net/publication/230667951_Local_dielectric_environment_of_B-DNA_in_solution_Results_from_a_14_ns_molecular_dynamics_trajectory?ev=prf_pub

Article Local Dielectric Environment of B-DNA in Solution: Results f...

Marksim and Matthew have good points. Think about what dielectric relaxation is actually doing in your polymer. Same thing as a physical relaxation test, only a different frequency. Think about it thermodynamically - energy into the system is energy in - whether strain energy of a physical test or electrical energy from a dielectric test. How does your MD simulation handle an energy input?

Energy in must either come back out again, or do work in some form or another. If you get a peak on a dielectric response spectrum, it is because the energy going in is doing work by moving molecular masses through distance by making them vibrate - that is the "loss" of the loss modulus trace. If your simulation has properly captured the mechanism of molecular vibration, you should be able to reproduce a dielectric response spectrum in an area of critical interest. Try a place on the spectrum of a known and well documented pure system first. Physical experimental systems might have low molecular weight oligomers that produce relaxation peaks in unusual places relative to a nice "pure" computational model. Make certain you have good baseline data to compare against.

system's collective dipole time correlation function is related to dielectric relaxation. It's Fourier transform is the system's frequency dependent susceptibility.