I've been making molds for a while and casting polyurethane and what they call white metal.
One of the silicone rubbers that i use is transparent and can handle up to 220 degrees celcius (i use it to bake polymer clay) and there are plenty silicone rubbers that can handle even higher temperatures.
For a simple limited series I just take my part (3d printed) and glue some pins to it that will make my pouring channels in the mold and then pour rubber arround it.
then i cut the mold open and remove the pins and part.
Frankly I cannot see any easier or cheaper way to make molds.
For more complex shapes i make multiple part molds which takes more work, but still i cannot imagine a easier or cheaper way to make molds.
The lost wax method usualy involves a one-off piece where the mold is destroyed to free the endproduct.
I think for the stainless steel prints they use that method.
You can use the same method you use for polyurethane to make metal parts in
woods metal (that is a compound with metal particles that cures at low temperature) which conducts electricity and is as strong as alluminium.
It can be used to pour circuitry inside a plastic part for instance and you can solder it.
There are some interesting articles about that stuff published by the university of Bath (UK)
Have a look at the RepRap site, there you can find the links in the Darwin version 2 articles.
Those guys are looking into 3dprinting electronic circuitry inside plastic parts in 1 run for their next version of the Reprap.
i believe that stuff cures at 70 degrees celcius or thereabouts.
Might be usefull for casting metal parts for your trainsets at low temperature.
Any silicone rubber can handle those temperatures since many plastics are hardned by heating them at those temperatures, for instance epoxy is post cured at those temperatures for extra strength.
And the mixing is not realy an argument anyone can weigh one part of the compound then weigh the other.
as far as airbubbles are concerned, If you mix it carefully (not to fast to prevent mixing in air) you get hardly any bubbles with modern RTV's.
For a fast mold i usualy find it sufficient to just burst airbubbles on the surface with a toothpick for 10 minutes after pouring the rubber, i rarely find bubbles pressed against the master, because i first apply a thin layer of rubber with a brush and let it cure first., then fill the moldframe
It's not fast, but it eliminates the need for a vaccuumchamber.
But then again i doubt there is any need for speed in the model train hobby