A fixed point is a value that maps to itself, x* = f(x*), so substituting it back in leaves the state unchanged. It is the discrete-map analog of an equilibrium: once there, the system stays. The exact closure holds only for the stated map; real systems add noise, parameter drift, and model error, so a computed fixed point is only as trustworthy as the map behind it.

A fixed point is a checked finite substitution. Exact arithmetic here means exact results for the stated model inputs; measured inputs still carry uncertainty and significant-figure limits.

highlighted = computed this step

Fixed point means the update returns itself

The candidate one half is checked by substitution.

f(12)=12⋅12+14=12f\left(\frac{1}{2}\right)=\frac{1}{2}\cdot\frac{1}{2}+\frac{1}{4}=\frac{1}{2}
Fixed-point substitutionThe output label must equal the computed update.mapId=affine_contractionparameter=nonecandidate=1/2outputValue=1/2fixedBit=1 bit

The fixed bit is computed, not chosen

Because the output is again one half, the fixed bit is 1.

f(x∗)=x∗⇒1f(x^*)=x^*\Rightarrow 1
Checked fixed bitA stale fixed-bit label fails validation.mapId=affine_contractionparameter=nonecandidate=1/2outputValue=1/2fixedBit=1 bit