Incidentally, "triangulation" as a DF or interference-hunting methodology is, in my experience, both very overrated and very misunderstood. Unless you are in a more or less reflection (multipath) free environment, it's very difficult to get good bearings: calculating the interception of those bearings is trivial, but junk bearings yield junk results.
It becomes more difficult at higher frequencies where reflections become stronger. Roanoke style Doppler direction finding systems have a lot more trouble than systems which rely on a beam antenna, but the sidelobes on beam antennas create their own problems. Even in challenging environments, I had excellent results with beam antennas designed to have a minimum of sidelobes, at which point it was possible to see multipath from things like trees which otherwise was concealed by stronger sources. With a really good antenna, one can "see" the terrain from its reflections. Once I had a good enough directional antenna, I could track flights into and out of LAX from their reflections as a form of poor man's bistatic radar.
Since interference / jamming happens around (or at least is important around) people, this means most practical DF'ing is being done in urban and suburban environments where multipath can be an issue.
That was definitely my experience. Large buildings with flat surfaces make for a challenging environment.
For cellular network operators, they know from the base station stats (RSSI, e.g.) which sector or sectors are being affected, and in most cases you can simply drive (or walk) the sector until you get close and have to hunt on foot (where triangulation is completely useless).
Triangulation is not always necessary. Once a direct path at close range is available, a rise in signal strength of 6dB indicates that the distance has halved, making an estimate of distance to the target possible. This of course depends on having a beam antenna with low sidelobes to reject multipath interference.
The 6dB rule applies equally well to reflections, so at close range it may be necessary to back off and find a higher location to try and lock onto a direct path signal.
Just wonder is it possible to use some kind of synchronized array of transmitters distributed all around the base stations and transmitted signals in that way so their interference signal at base station location gives false position of the transmitter or just spamming it with a bunch of virtual transmitter positions to make it hard to find which one is real? Are such kind of jammers are used in practice?
There are a lot of "creative" approaches to jamming, but high-end direction finding systems (like the ones we make) can usually still DF people trying to use "creative" techniques. Generally speaking, for a jammer to be effective it has to be (a) loud, (b) wide, and (c) on, and all three of those things make jammers relatively easy to DF, regardless of how they are implemented. A weak signal with a low duty cycle that's only a few kHz wide is harder to DF, but it's also not an effective jammer.
I tried something once on one of our amateur band transmitter hunts which was somewhat effective. We hunted standard 2 meter FM signals, but knowing how signal strength and noise meters responded, I modulated the amplitude with low frequency noise through an exponential circuit to create a linear response on various signal strength meters. This made it more difficult for hunters who relied on signal strength to gain a bearing, but had no effect on Doppler direction finders of course.