Definition

Let π1:GU(H1)\pi_1:G\to U(H_1) and π2:GU(H2)\pi_2:G\to U(H_2) be of the same . They are unitarily equivalent if there is a U:H1H2U:H_1\to H_2 such that

Uπ1(g)=π2(g)Ufor every gG.U\pi_1(g)=\pi_2(g)U\qquad\text{for every }g\in G.

Thus a unitary equivalence is an invertible that also preserves . It identifies the two actions after an isometric change of Hilbert-space coordinates, rather than asserting that their operators are literally equal on one fixed space.

Invariants

A unitary equivalence carries closed invariant subspaces of H1H_1 to closed invariant subspaces of H2H_2. It therefore preserves irreducibility, orthogonal direct-sum decompositions, multiplicities, and after vectors are transported by UU. The integrated representations also have the same and corresponding kernels.

Equivalence relation and classes

Identity operators, adjoints, and compositions show that unitary equivalence is an . Representation theory normally classifies unitary representations up to this relation. In particular, the of a consists of unitary-equivalence classes of ; see Folland, §3.1.

Conventions and scope

For unitary representations, a bounded invertible intertwiner does imply unitary equivalence: the unitary factor in its polar decomposition still intertwines the actions. The original intertwiner itself need not be unitary. Equality of characters or of selected coefficient functions does not, without an appropriate classification theorem, constitute the definition of unitary equivalence.

References
  1. Gerald B. Folland, A Course in Abstract Harmonic Analysis, 2nd ed., CRC Press, 2016. DOI record. Relevant: §3.1 on unitary representations, intertwiners, and equivalence.