Do you want to do it dynamically in analog, or on paper?
For paper, I often use Excel to figure what a thermistor will do over a temperature range, and of course I can make a graph of the function, and if I set it up right I can have it give me the equation of the solution (temperature out from ADC counts in).
I've used some gas-meter gauge transducers and done likewise. You don't need their equation, although you can have Excel calculate it for you, just the measured points (A = X counts, B = Y counts, C = Z counts, etc). And you can figure the equation for the total conversion, to at least sixth power with Excel 2003 and later. (Excel 2000 really stops at 4th, even though it SAYS 6th. With simple thermistor connections, I have had the best ADC count to temperature curve-matching luck with exponentials.
For analog, yeah diode breaks and fet curves and resistors and such can do it. Temperature can be an issue. Temperature can also affect the gauge transducer, and the gauge itself.
<edit>Oops, I just re-read what you wrote. There are some electronically-controlled potentiometers and rheostats, if the strain gauge current or voltage are not too high.
For higher power stuff, if applied voltage is unipolar, you can make a controllable current load, and drive it from the output of a DAC, and adjust current based on the applied voltage, and emulate pretty much any E-I characteristic. The problem with this is that a simple variable-current load drops a volt or two minimum. If that is tolerable, then you are OK, and the computer can work out the curve, either by equation or by a point-by-point lookup table. You may need to work temperature compensation into it.
I do something like that when testing solar panels. I built a variable current load and I measure panel voltage versus load current, and I use the computer to vary the load current while reading voltage and current. I drive this with a "LabJack" I/O unit. Works good. The LabJack I use has a number of simultaneous ADC and DAC ports so I can monitor ambient temperature, and a reference panel short-circuit current, while I run the panel under test through its range. If you're interested, I'll grab the model number. <EDIT> What I have is the LabJack U12. There are other models as well.
www.labjack.comA more sophisticated variable-current load can operate with very little voltage drop. Or, you can "ground" the variable-current load to a negative voltage, so it works all the way to ground, and you stop there.
