Tantalos said:
I am not assuming that one of the twins is stationary, but that one is moving with a different speed than the other. But this situation is symmetric, when A is moving away from B, then also B is moving away from A. Both can make the calculation of their ages and A comes to the conclusion that B will be younger, then B will make the calculation and to him A should be younger.
Like I said, this is true if the twins always maintain a single velocity.
But if you are going to use the "v" in the Lorentz transform to apply to the difference in speed between the twins, even if they are both "moving" (whatever that means), then don't you mean that in the two reference frames, one of them is stationary and the other is moving? Of course you don't have to do this but if you don't, you will have to be more precise in what you mean because there are an infinite number of ways to interpret what you are talking about.
Tantalos said:
When the twin wants to return then his speed changes to -v. Then we cannot make a new calculation because their origins of their reference frames are not in the same place.
Only if we don't put them in the same place. Just like we don't have to presume that an observer/object/clock is stationary in a Frame of Reference, we don't have to put any observer/object/clock at the origins and we certainly don't have to put both observers/objects/clocks at the same origin.
So if we make the origins of the two Frames of Reference be the turn-around event for the traveling twin, then we won't introduce any spurious offsets in the calculation of his aging during his trip.
But we don't even have to do that, we can continue to use the same Frame of Reference in which the "traveling" twin was at rest during the outbound portion of his trip and during which the other twin was actually traveling away from him at some constant speed "v" throughout the entire scenario. Then at the turn-around event, the "traveling" twin has to go from rest to a speed greater than "-v" in order to catch up to the other twin.
It doesn't matter which Frame of Reference we use to analyze the entire scenario of the Twin Paradox. The easiest one to use is the one they both start out at rest and in which only the traveling twin has a non-zero speed because the calculation is trivially simple but if you like to torture yourself, you can pick a different Frame of Reference and endure the more complicated calculations but you will end up with the same age difference in the two twins between the time they separate and the time they reunite.
Tantalos said:
Einstein is assuming that the reference frame can change its speed or direction and the Lorentz transformation still can be used. Which is true it can or cannot?
It cannot according to Einstein's formulation of a Frame of Reference in his
1905 paper and he doesn't make that assumption. Where did you get that idea?
In fact, if you look at the end of section 4 of his paper, you will see that he introduces the Twin Paradox and its solution to the world for the first time, although he doesn't call it that and he doesn't use twins, he uses a pair of clocks. And he assigns a single Frame of Reference to both clocks which he calls K, the stationary system.