= Skew product
{title2=$F(x,y)=(R(x),S_x(y))$}
A <skew product> evolves a base coordinate independently while allowing a fiber coordinate to depend on the base:
$$
F(x,y)=(R(x),S_x(y)).
$$
The first coordinate obeys the autonomous rule $x_{n+1}=R(x_n)$, while $y_{n+1}=S_{x_n}(y_n)$ is driven by the base state. Continuity, smoothness or measurability is imposed according to the setting.
For example, $F(\theta,y)=(\theta+\omega\pmod1,ay+\sin(2\pi\theta))$ describes a linearly contracting response when $|a|<1$, driven by a circle rotation. Varying $y$ leaves every future base angle unchanged. By contrast, a coupled map whose first coordinate also depends on $y$ is not a skew product in these coordinates.
An <invariant graph of a skew product> satisfies $h(R(x))=S_x(h(x))$: starting at $(x,h(x))$ lands at $(R(x),h(R(x)))$. The <fiber Lyapunov exponent> measures response to perturbations within a fiber, separately from growth in the base. This separation is useful for <strange nonchaotic attractors>.
The <irrational skew shift> gives an affine example on a <torus>, while the <circle skew-product minimality criterion> identifies when a circle extension of a minimal base has no proper closed invariant subsystem.
Back to article page