Frequency-variable impedance in a lumped port

Hello everyone,
I am struggling to design a matching network by pieces and I would like to introduce a lumped port whose impedance depends on frequency. I have not been able to find an efficient way of doing it and I am getting some awkward results just by using impedance average values over the band.

Does anyone know how to get this done?

It depends on port impedances to be matched. Wideband matching can't be achieved in any case, there are some principal limitations for lossless passive networks. E.g. you can't reduce the Q of a resonant circuit by passive elements (neither lumped nor distributed).

Thank you, FvM. However, my question is a different one: is it possible to introduce a frequency-variable impedance into a lumped port, just as you would do with a N-port network in Ansoft Designer, for instance, where you could use the S matrix of a s2p file among other formats. The thing is: I am to going to plug my design into an antenna whose impedance profile over the band is not constant, as you can expect, and it is not good enough to assign a fixed valued impedance at the end port.

So it's either the question if a particular tool is supporting S parameter tables for a port specification or how to design a lumped element approximation for an empirical impedance characteristic. The former should be answered from the respective manuals, the latter is possible by any tool, that can calculate circuit impedances and perform a parameter fit. Of course you have to provide a suitable circuit model. For an antenna, a RLC series circuit would be promising as a first guess. More general, even a spreadsheet calculator like Excel can perform the fit, if you give it a formula for the circuit impedance.

Yes, sorry, I thought I had said it in the first place, but I hadn't. I am trying to do this with HFSS. Finding a fit might be a solution, but I was wondering if I can import the S11 or the Input Impedance straight into the port.