A left Haar measure on a locally compact group that is a Hausdorff space is a nonzero regular Borel measure, finite on compact sets, satisfying for every Borel set . Such a measure exists and is unique up to a positive scalar. Right Haar measure instead satisfies ; the two need not coincide. On a compact group, normalization gives total mass one and invariance on both sides. See Tornier, Haar Measures, Definition 2.1 and Theorem 2.2.
For a finite group, normalized Haar measure is uniform counting; for the circle group, it is normalized angle integration on the circle. These examples turn symmetry into a rule for averaging.
Generate a real -by- matrix with independent entries having the standard normal distribution. Take its QR decomposition with positive diagonal in the upper triangular matrix factor, . The orthogonal matrix has normalized Haar measure on the full orthogonal group . If a numerical factorization returns , set and return . The signs are nonzero almost surely: successively drawn Gaussian columns have probability zero of lying in the proper linear subspace spanned by earlier columns.
For correctness, every fixed orthogonal matrix preserves the standard Gaussian column density and column independence, so and have the same probability distribution. The positive-diagonal QR decomposition is unique: two decompositions would give an orthogonal matrix that is also an upper triangular matrix with positive diagonal, hence the identity. Therefore and has a left-invariant probability law. Uniqueness of normalized Haar measure on the compact group identifies that law. Each determinant sign has probability by reflection invariance.
On a -dimensional Lie group, take the coframe dual to a basis of left-invariant vector fields. Its top wedge is nonvanishing and invariant under left translations, so its absolute density defines a left Haar measure. If the frame has coordinate matrix , the density is . This makes explicit computation of Haar measure a determinant calculation.
For nonzero nonnegative , let be the infimum of over finite dominations with . Composition of covers gives , so is uniformly bounded while . It is invariant under left translation, homogeneous, monotone and subadditive. As the support of shrinks to the identity, its additivity defect tends to zero: use the continuous proportions , with a cutoff positive on both supports, whose oscillations are small on every translate of that small support. In a locally compact sigma-compact metrizable group, diagonal convergence on a countable dense family, followed by uniform approximation on fixed compact supports, gives a positive invariant linear functional on all . The Riesz representation on compactly supported continuous functions turns it into Haar measure. Local compactness is essential for the compactly supported cutoffs and small-support test functions.
For a compact group , positive normalized linear functionals on form a nonempty weak-star topology compact convex set. For finitely many translations, apply the Schauder-Tychonoff fixed-point theorem to the average of their induced affine dual maps. A fixed functional is stationary under this average. For each real continuous , its orbit function is continuous on the compact subgroup generated by those translations and satisfies a finite averaging identity. At a maximum, every summand must attain that same maximum; iteration and density of the positive-word semigroup force to be constant. Thus the functional is fixed by each of the finitely many translations. Compactness and the finite intersection property now give a functional invariant under every translation. The Riesz-Markov-Kakutani representation theorem turns it into a normalized Haar measure, without assuming commutativity of the group.
A Haar integral integrates a function against Haar measure. On a compact group, normalizing the measure to total mass one makes it an averaging operation. Translation invariance then implies that the average of a nontrivial continuous unitary character is zero: translating the integral multiplies it by a character value different from one.
For a class function on SU(2) and normalized Haar measure, the displayed formula reduces a group integral to its conjugacy-class angle. The weight follows by identifying SU(2) with the unit sphere in four real dimensions: writing a unit quaternion as gives spherical volume proportional to . Since , normalization gives .
The hemisphere graph of SU(2) as the three-sphere has induced metric , whose determinant is . Its volume density is therefore . Fixed left or right multiplication preserves the Euclidean norm on the four Pauli coefficients, so the round volume is Haar measure. Both hemispheres have volume , giving total volume and normalized density on each hemisphere. The apparent equatorial divergence is an integrable coordinate singularity.
For an open identity neighbourhood in a compact topological group, a finite left covering net is a finite set such that the left translates , , cover the group. Compactness gives such finite sets; minimum cardinality gives a minimal one. Left translation preserves minimality because implies .
Two minimal left covering nets for the same identity neighbourhood can be matched bijectively so that paired translates intersect. Join to when . For , its neighbours cover every point covered by , so still covers the group. Minimality forces . Hall's marriage theorem supplies the matching, and an intersection gives . Averaging a continuous function over matched nets therefore gives values differing by at most its uniform oscillation on right displacements in . This is the finite combinatorial step in constructing Haar measure on a compact group.
Articles by others on the same topic
The Haar measure is an important concept in the area of harmonic analysis and abstract algebra, specifically in the context of topological groups. It is a way of defining a measure on a locally compact topological group that is left-invariant (or right-invariant), which means it remains unchanged (invariant) under the group's operations.