Eigenspace invariance for commuting linear maps Created 2026-10-25 Updated 2026-10-31
If two linear maps commute and , then . Hence preserves each eigenspace of . On a one-dimensional eigenspace, is a scalar multiple of , so each eigenvector for a simple eigenvalue of is also an eigenvector of . For repeated eigenvalues the invariant eigenspace can have higher dimension; arbitrary vectors in it need not individually be eigenvectors of .
Eigenvalue sensitivity Created 2026-09-24 Updated 2026-11-05
For unit left eigenvector and right eigenvector of a real matrix at a simple eigenvalue , its normwise sensitivity isIt measures the first-order amplification of a matrix perturbation into an eigenvalue perturbation.
Let be fixed real symmetric matrices, with , , and a simple eigenvalue. Complete to an orthonormal eigenbasis with eigenvalues . Put and . Then . The symmetric eigenvalue residual bound produces an eigenvalue in Big O notation.
Irreducible nonnegative matrix Created 2026-10-05 Updated 2026-11-05
A square nonnegative matrix is irreducible if for each pair there is an integer with . This says that every index can reach every other through positive entries. The Perron–Frobenius theorem then gives a positive leading eigenvector and an algebraically simple eigenvalue. Irreducibility alone permits other eigenvalues of the same modulus: the cyclic permutation matrix has eigenvalues .
Past exam of the mathematics course of the University of Cambridge 2017 ii Paper 4 16H Solution 2026-10-05
For the Petersen graph, each vertex has three neighbours, adjacent vertices have no common neighbour, and distinct nonadjacent vertices have exactly one common neighbour. The entries of count length-two walks, so diagonal entries are three, adjacent entries zero, and other off-diagonal entries one. HenceThe connected graph has the simple eigenvalue three with eigenvector . On its orthogonal complement , so the remaining eigenvalues satisfy and are one or minus two. If their multiplicities are , then and . Thus
Past exam of the mathematics course of the University of Cambridge 2023 ii Paper 4 40C b Solution Created 2026-09-24 Updated 2026-09-25
Differentiate the eigenvalue equationat , where . Suppressing the argument givesLeft-multiplication by the transpose of the unit left eigenvector and use of cancel the terms involving :Because the eigenvalue is simple, , and hence the first-order perturbation of a simple eigenvalue isThe definition of the operator norm, followed by the Cauchy-Schwarz inequality, givesTherefore
Past exam of the mathematics course of the University of Cambridge 2026 ii Paper 2 34B Solution Created 2026-09-24 Updated 2026-10-03
If and , differentiation and the Lax equation give , so after projection onto the normalized eigenfunction for a nondegenerate eigenvalue.
For ,The compatibility is precisely . If all coefficients are -independent, this becomes , and cyclicity of trace gives .
Past exam of the mathematics course of the year 2 of the natural sciences course of the University of Cambridge 2003 Paper 1 5 b Solution Created 2026-10-18 Updated 2026-10-20
Let . The commutation relation gives , so preserves the eigenspace of with eigenvalue . Because it is a nondegenerate eigenvalue, the eigenspace is one-dimensional, and for some real . This includes . Thus the eigenvectors of form a common eigenbasis, andWriting the corresponding unit-modulus eigenvalue of as givesThe pairing refers to the same common eigenvector; independent reorderings of the two real spectra would lose this information. The simultaneous diagonalization can also be constructed without requiring simple eigenvalues by diagonalizing within each eigenspace of .
Choose a shift which is not an eigenvalue. Inverse iteration applies power iteration to , implemented by a linear system solve rather than by forming an explicit matrix inverse:A single LU decomposition of can be reused for all the solves at this fixed shift. An eigenpair transforms intoThus an eigenvalue uniquely closest to becomes the dominant eigenvalue of . For a diagonalizable matrix, the argument in the preceding subpart gives convergence of diagonalizable fixed-shift inverse iteration, provided the starting coefficient in the target eigenspace is nonzero, with power iteration direction error , whereFor any prescribed simple eigenvalue, taking a sufficiently close but unequal shift makes the numerator small while the denominator stays close to its positive separation from the remaining eigenvalues. This makes small and gives very fast convergence per iteration. The Rayleigh quotient of estimates directly. If estimates the associated eigenvalue of , the corresponding estimate is .
The selection principle also applies when the other eigenvalues have nontrivial blocks in the Jordan normal form: their contributions acquire powers of , multiplying strictly smaller geometric factors. A simple uniquely closest eigenvalue still attracts the direction when the Riesz projection of onto its generalized eigenspace is nonzero. For nonreal target eigenvalues, allow complex shifts and arithmetic; a real shift cannot distinguish a conjugate pair of equal distances. Setting makes the solve singular, and an exactly zero starting target component prevents convergence to that target in exact arithmetic. When is real, account for the sign behavior of fixed-shift inverse iteration.
Perron–Frobenius theorem 2026-10-05
The strictly positive case states that a real square matrix with has a positive simple eigenvalue and positive left and right eigenvectors, with all other eigenvalues strictly smaller in modulus. The Brouwer fixed-point theorem applied to on the nonnegative unit simplex gives and . For another eigenvector , maximize ; the triangle inequality gives . Equality forces every modulus ratio and complex argument to agree because every entry is positive, so is proportional to and . Apply the same argument to for a positive left eigenvector; its positive pairing with rules out a generalized eigenvector at , establishing algebraic simplicity. For merely nonnegative matrices a nonnegative leading eigenvector exists. An irreducible nonnegative matrix has a positive leading eigenvector and simple Perron eigenvalue; a primitive nonnegative matrix has the strict modulus gap.
Suppose is a diagonalizable matrix, , with an eigenbasis and spectral decompositionAssume the dominant eigenvalue is a simple eigenvalue andThen the normalized power iteration approaches the one-dimensional eigenspace of . Its asymptotic direction-error bound has factor . This spectral gap is a gap in modulus: if , then . A small relative gap gives slow convergence.[ref]
The short proof is the exact eigenbasis identityThe remaining sum has norm bounded by a constant times . This proves convergence of direction, allowing a changing sign or complex factor. A tiny or a poorly conditioned, nonorthogonal eigenbasis can delay visible convergence: the condition number of affects constants, while controls the eventual rate. The bound need not be an equality; if the initial coefficient of the second-largest eigenvalue vanishes, the active spectrum of the power method gives a faster rate. When , all unwanted components disappear after one multiplication.[ref]
The hypothesis is , not an ordinary nonzero inner product with the right eigenvector. Let be the first row of , a left eigenvector normalized so is for and otherwise; denotes conjugate transpose. Then . For a Hermitian matrix, choose an orthonormal eigenbasis and the criterion becomes : the starting vector must be nonorthogonal to the dominant eigenvector. For a general diagonalizable matrix, the nonorthogonal eigenbasis makes these two tests different, as illustrated under power iteration failure cases.
Suppose a unitary matrix is symmetric under ordinary transpose, and write its entrywise real and imaginary parts as and . They are real symmetric matrices. Expanding gives , proving the displayed identities. Commuting real symmetric matrices admit simultaneous diagonalization: diagonalize , then diagonalize the restriction of to each invariant eigenspace of . A common real orthonormal eigenbasis givesIf has simple eigenvalues, every eigenvector of is already an eigenvector of , because its one-dimensional eigenspace is invariant. The eigenvalues of and are paired as the real and imaginary parts of the same eigenvalue of .