it's more like educated trial and error.
Let's do the first one to demonstrate why it's not just random trials.
We have 4 in the denominator so it's unlikely the denominator of the addends will be 3, 5, or 6
simple ways of making 5/4 might be 1 + 1/4, 1/2 + 3/4, 2 - 3/4 etc
well we can't do 1+ 1/4 as we repeat the 1
we've got no way to get the 3/4 negative
but what we can do is 3/4 + (-1)/(-2) = 5/4
So with a little thought we knocked down that 6! to a quick check of 3 likely cases.
I suspect the remaining two will be similar.