Hmm, I think that the range would be [0, infinity
) (making sure to use the
round bracket when using -infinity or infinity in your domain

).
Why? The way I like to look at it, is to analyse the motion in the x-axis (controlled by the
i component) and the motion in the y-axis (controlled by the
j component) separately.
For your motion in the x-axis, your particle basically oscillates from 2 to -2 and back again, much like an ideal spring would. So it'll just go back and forth repeatedly, achieving its maximum speed at 0, and its minimum speed at 2 and -2, due to the effect of the 'cos' in the expression of
i. For your motion in the y-axis, your particle just travels upwards at a rate of 3 units per unit of time. But there is no reason at all why the particle has to stop after going from 2, to -2 and back to 2 again. So there isn't really a restriction on how 'positive' the range can be.
I like nliu1995's approach to this question; finding x in terms of y eliminates a lot of hassle. With regard to your query though, MathsQuest says cos^-1(x) and Cos^-1(x) are the same thing and Wikipedia says they are merely separate conventions. I honestly can't remember what is required for VCAA purposes, (which is probably what is most important at this point in time), but it seems they say cos^-1(x) already has the restricted domain in their formula sheet.