A differential equation system is given by: for all t Î Reals+
z'(t) = g(z(t), v(t)).
Here
z'(t) is the derivative of z with respect to t, evaluated at t ;
t Î Reals+ stands for continuous time;
z: Reals+ ® Reals N is the state response; and
v: Reals+® Reals N is the input signal.
Here, the state is updated continuously rather than discretely, but otherwise, this is similar to a state machine with an infinite state space. In fact, we can approximate the derivative by the difference
z(t + d ) - z(t) @ d g(z(t), v(t)).
If we sample z and v at times 0, d , 2d , ..., nd, (n+1)d , ... and denote the nth sample value by
s(n) = z(nd ), x(n) = v(nd ),
we get the difference equation
s(n + 1) - s(n) = d g(s(n), x(n))
This is called a difference equation because the difference on the left is analogous to a differential in a differential equation. This can be rewritten as a state update equation,
s(n + 1) = s(n) + d g(s(n), x(n))
Thus, a discrete approximation to a differential equation is a state machine with an infinite state space.