A linear operator that pushes densities forward under a dynamical map, describing how probability distributions evolve by transporting mass according to the map's inverse Jacobian.

Principle

For a measurable transformation, the Perron–Frobenius operator L acts on measures or density functions ρ so that ∫_A Lρ equals ∫_{f^{-1}(A)} ρ; it encodes the statistical evolution induced by the map.

Demonstration

For the map on the unit interval f(x)=2x mod 1, the operator leaves the uniform density invariant; computing L shows that L1 = 1, indicating preservation of Lebesgue measure.

Misapplication

Applying the operator formalism without checking absolute continuity (acting on densities when the invariant measure is singular) yields misleading conclusions about mixing and relaxation.

Consequence

Used correctly, it yields invariant densities, rates of mixing via spectral gaps, and allows prediction of long-term statistical properties from the map's action.

Reversal

Instead of pushing densities forward, the adjoint action (Koopman operator) pulls observables forward; contrasting these two perspectives clarifies whether one focuses on states (densities) or measurements (observables).

Boundary

Defined for measurable maps and requires a choice of function space (L1, BV, etc.); spectral properties depend on that space and on regularity assumptions of the map (expanding, piecewise smooth).

Semantic Tension

Competes with the Koopman operator viewpoint; both encode the same dynamics but on dual spaces (densities vs. observables), which can lead to different analytical tools and intuitions.

Synthesis

The Perron–Frobenius operator is the linear transport operator on densities induced by a dynamical map that encodes invariant measures and statistical evolution through its spectral structure.