Almaz Khusnutdinov•Math enthusiast, love coding and mathematics
Published 11/2/2025
Triangular Forms
Problem 1.1
Let A=200130045.(a) Compute Δ(t)=det(tI−A).(b) Find the eigenvalues and verify that tr(A),det(A) equal the sum/product of eigenvalues.Detailed solution: tI−A=t−200−1t−300−4t−5is upper triangular, soΔ(t)=det(tI−A)=(t−2)(t−3)(t−5).Eigenvalues: 2,3,5.tr(A)=2+3+5=10,det(A)=2⋅3⋅5=30.Sum/product of eigenvalues: 2+3+5=10,2⋅3⋅5=30.✓Answer: Δ(t)=(t−2)(t−3)(t−5),eigenvalues 2,3,5,tr(A)=10,det(A)=30.
Problem 1.2
Prove (outline): If A is triangular, then Δ(t)=det(tI−A)=i=1∏n(t−aii).Answer (outline): (1) tI−Ais triangular with diagonal entries t−a11,…,t−ann.(2) The determinant of a triangular matrix is the product of its diagonal entries.(3) Hence det(tI−A)=(t−a11)⋯(t−ann).
Problem 1.3
Prove (outline): If the characteristic polynomial of T:V→Vsplits over F,then there exists a basis in which [T]is upper triangular.Answer (outline): (1) Splitting ⇒there exists an eigenvalue λand eigenvector v1.(2) Extend v1to a basis B=(v1,…,vn).In this basis, [T]B=[λ0∗∗](block upper triangular).(3) The span of v1⊥(the remaining n−1 vectors) is T-invariant; restrict Tto it.(4) Apply induction on dimension to triangularize the restriction; combine blocks to get an upper triangular [T].
Problem 1.4
Let A=100110002.Find the minimal polynomial mA(t).Answer: mA(t)=(t−1)2(t−2).
Problem 1.5
Let A=a00∗b0∗∗cbe upper triangular (stars arbitrary), and let m∈N.Compute det(Am)and tr(Am).Answer: det(Am)=(abc)m,tr(Am)=am+bm+cm.
Problem 1.6
Consider the following upper triangular matrices.(a) A=10023004−1.Compute the characteristic polynomial ΔA(t) and the eigenvalues of A.(b) B=200020135.Compute the characteristic polynomial ΔB(t) and the eigenvalues of B.Detailed solution:(a) For A,tI−A=t−100−2t−300−4t+1.This matrix is upper triangular, soΔA(t)=det(tI−A)=(t−1)(t−3)(t+1).Hence the eigenvalues of A are 1,3,−1.(b) For B,tI−B=t−2000t−20−1−3t−5,which is also upper triangular.ΔB(t)=det(tI−B)=(t−2)(t−2)(t−5)=(t−2)2(t−5).Thus the eigenvalues of B are 2 (with algebraic multiplicity 2) and 5.Answer: ΔA(t)=(t−1)(t−3)(t+1),λA∈{1,3,−1};ΔB(t)=(t−2)2(t−5),λB∈{2,2,5}.
Problem 1.7
Let C=400−140024,D=00015000−2.(a) Compute the characteristic polynomial and eigenvalues of C.(b) Compute the characteristic polynomial and eigenvalues of D.Answer: ΔC(t)=(t−4)3 with eigenvalue 4 of algebraic multiplicity 3;ΔD(t)=t(t−5)(t+2) with eigenvalues 0,5,−2.
Problem 1.8
Consider the triangular matricesA=1001−20034,B=[−1023].(a) For A,find the eigenvalues of A and of A2 using triangular form.(b) For B,find the eigenvalues of B and of B2 using triangular form.Answer: For A:λ(A)={1,−2,4},λ(A2)={1,4,16}.For B:λ(B)={−1,3},λ(B2)={1,9}.
Problem 1.9
Consider the upper triangular 2×2 matrices(a) A=[3013],(b) B=[a02b],a,b∈C.For each matrix, compute the characteristic polynomial and list its eigenvalues (with multiplicities when appropriate).Answer: ΔA(t)=(t−3)2 with eigenvalue 3 of algebraic multiplicity 2;ΔB(t)=(t−a)(t−b) with eigenvalues a and b.
Problem 1.10
Let A be a square matrix over C.(a) Suppose 4 is an eigenvalue of A2.List all possible eigenvalues μ of A such that μ2=4.(b) Suppose 91 is an eigenvalue of A2.List all possible eigenvalues μ of A such that μ2=91.Answer: (a) μ=2 or μ=−2;(b) μ=31 or μ=−31.
Problem 1.11
Consider the following upper triangular matrices.(a) A=10002100034000−14.Compute the characteristic polynomial of A and list the eigenvalues of A2 (with multiplicities).(b) B=200−12005−3.Compute the characteristic polynomial of B2 and list the eigenvalues of B2 (with multiplicities).Answer: ΔA(t)=(t−1)2(t−4)2,λ(A2)={1,1,16,16};ΔB2(t)=(t−4)2(t−9),λ(B2)={4,4,9}.
Problem 1.12
Let A,B be 3×3 complex matrices whose characteristic polynomials factor into linear polynomials.(a) Suppose the eigenvalues of A2 are 1,9,9 (counted with multiplicity) and tr(A)=5.(a)Find the eigenvalues of A (with multiplicities).(b) Suppose the eigenvalues of B2 are 4,4,25,det(B)=−20,tr(B)=5.(b)Find the eigenvalues of B (with multiplicities).Answer: (a) eigenvalues of A:−1,3,3;(b) eigenvalues of B:−2,2,5.
Invariance
Problem 2.1
Let A,B be linear operators on R3 with matricesA=010−100002,B=100020003.(a) Find all subspaces of R3 that are invariant under A.(b) Find all 1-dimensional and all 2-dimensional subspaces of R3 that are invariant under B.Answer:(a) Invariant subspaces under A:{0},R3,span{(0,0,1)},span{(1,0,0),(0,1,0)}.(b) Invariant subspaces under B:1-dimensional: span{(1,0,0)},span{(0,1,0)},span{(0,0,1)}.2-dimensional: span{(1,0,0),(0,1,0)},span{(1,0,0),(0,0,1)},span{(0,1,0),(0,0,1)}.
Nilpotent Operators
Problem 3.1
Let dimV=7 and T:V→V be nilpotent. Suppose dim(kerT)=3,dim(kerT2)=5,dim(kerT3)=6,dim(kerT4)=7.(a) Find the index k of T.(b) Determine the sizes of the Jordan nilpotent blocks of T and write a Jordan matrix for T.Solution: (a) Since dim(kerT4)=7=dimV, we have T4=0 so k≤4.Also dim(kerT3)=6<7, so T3=0. Hence k=4.(b) For a nilpotent Jordan block Ns of size s, one has dim(kerNsi)=min(i,s)⇒dim(kerNsi)−dim(kerNsi−1)={1,0,i≤s,i>s.Let ni=dim(kerTi) with n0=0. Define ai=ni−ni−1.Then ai equals the number of Jordan blocks with size ≥i.n0=0,n1=3,n2=5,n3=6,n4=7⇒a1=3,a2=2,a3=1,a4=1,a5=0.Now the number of blocks of exact size i is ai−ai+1:#(4-blocks)=a4−a5=1,#(3-blocks)=a3−a4=0,#(2-blocks)=a2−a3=1,#(1-blocks)=a1−a2=1.So the block sizes are 4,2,1 (they sum to 7), and the total number of blocks is a1=3=dim(kerT).Answer: k=4,Jordan block sizes (4,2,1),[T]J=diag0000100001000010,[0010],[0].
Problem 3.2
Let dimV=6 and T:V→V be nilpotent with dim(kerT)=2,dim(kerT2)=4,dim(kerT3)=6.Find the index k and the Jordan nilpotent block sizes of T.Answer: k=3,block sizes (3,3),[T]J=diag000100010,000100010.
Problem 3.3
Let dimV=8 and T:V→V be nilpotent with dim(kerT)=3and index k=4.List all possible multisets of Jordan nilpotent block sizes for T.Answer: {4,3,1}or{4,2,2}.
Problem 3.4
Let S,T:V→V be nilpotent operators on a 7-dimensional space V with dim(kerS)=3,dim(kerS2)=5,dim(kerS3)=6,dim(kerS4)=7,dim(kerT)=3,dim(kerT2)=5,dim(kerT3)=7.(a) Find the Jordan block sizes (and index) for S and for T.(b) Are S and T similar?Answer: S has index 4 and block sizes (4,2,1).T has index 3 and block sizes (3,3,1).S and T are not similar.
Problem 3.5
Let dimV=9 and T:V→V be nilpotent. Supposedim(kerT)=4,dim(kerT2)=7,dim(kerT3)=8,dim(kerT4)=9.(a) Find the index k of T.(b) Determine the Jordan nilpotent block sizes of T.Answer: k=4,block sizes (4,2,2,1).
Problem 3.6
Let T:R6→R6 be defined by the matrix (standard basis)A=000000100000010000000000000100000000.(a) Find the nullity dim(kerT) and the index k.(b) Give the Jordan nilpotent block sizes of T.Answer: dim(kerT)=3,k=3,block sizes (3,2,1).
Jordan Canonical Form
Problem 4.1
Let T:V→V be a linear operator withΔT(t)=(t−2)4(t+1)3,mT(t)=(t−2)3(t+1)2.(a) List all possible Jordan block size-multisets for the eigenvalues 2 and −1.(b) If additionally dimker(T−2I)=2 and dimker(T+I)=2, determine the Jordan form J of T.Solution: For an eigenvalue λ, let nλ be the exponent of (t−λ) in ΔT(t),and let mλ be the exponent of (t−λ) in mT(t).Then nλ equals the total size of the Jordan blocks for λ,and mλ equals the size of the largest Jordan block for λ.Also, the number of Jordan blocks for λ equals dimker(T−λI).Eigenvalue 2:n2=4,m2=3.So the Jordan block sizes for 2 form a partition of 4 whose largest part is 3.The only possibility is (3,1).Eigenvalue −1:n−1=3,m−1=2.So the Jordan block sizes for −1 form a partition of 3 whose largest part is 2.The only possibility is (2,1).Thus for (a): for 2 we must have (3,1),and for −1 we must have (2,1).For (b): dimker(T−2I)=2⇒there are exactly 2 Jordan blocks for 2,which matches the partition (3,1).dimker(T+I)=2⇒there are exactly 2 Jordan blocks for −1,which matches the partition (2,1).Therefore J=diag(J3(2),J1(2),J2(−1),J1(−1)).Answer: (a) For λ=2:(3,1) only;for λ=−1:(2,1) only.(b) J=diag200120012,[2],[−101−1],[−1].
Problem 4.2
Let T:V→V haveΔT(t)=(t−1)5(t−3)2,mT(t)=(t−1)2(t−3).(a) Is T diagonalizable?(b) List all possible Jordan block size-multisets for T.Answer: (a) No (since the exponent of (t−1) in mT is 2).(b) For λ=1:(2,2,1) or (2,1,1,1);for λ=3:(1,1).So J=diag(blocks for 1,I2⋅3) with the two possibilities above.
Problem 4.3
Let T:V→V haveΔT(t)=t6(t−4)3,mT(t)=t3(t−4)2.Find all possible values of dimkerT and the value of dimker(T−4I).Answer: dimker(T−4I)=2,dimkerT∈{2,3,4}.
Problem 4.4
Let T:V→V haveΔT(t)=(t−2)3(t+2)3,mT(t)=(t−2)2(t+2)3,dimker(T−2I)=2,dimker(T+2I)=1.Determine the Jordan block sizes for each eigenvalue and write a Jordan matrix J.Answer: λ=2:block sizes (2,1);λ=−2:block sizes (3).J=diag[2012],[2],−2001−2001−2.
Problem 4.5
Let T:V→V satisfyΔT(t)=(t−1)6,mT(t)=(t−1)4,dimker(T−I)=2,dimker((T−I)2)=4.Determine the Jordan block sizes of T and write a Jordan matrix J.Answer: block sizes (4,2),J=diag1000110001100011,[1011].
Rational Canonical Form
Problem 5.1
Let V be a vector space of dimension 7 over Q, and let T:V→V have minimal polynomialmT(t)=(t3−2)(t−1)2.(a) Determine the characteristic polynomial ΔT(t).(b) List all possible rational canonical forms M (as block diagonal matrices of companion blocks).Answer: ΔT(t)=(t3−2)(t−1)4.M=diag(C(t3−2),C((t−1)2),C((t−1)2))orM=diag(C(t3−2),C((t−1)2),C(t−1),C(t−1)).That is, C(t3−2)=010001200,C((t−1)2)=[01−12],C(t−1)=[1].
Problem 5.2
Let V be a vector space of dimension 8 over Q, and let T:V→V have minimal polynomialmT(t)=(t4+t+1)(t−2)2.(a) Determine the characteristic polynomial ΔT(t).(b) List all possible rational canonical forms M (as block diagonal matrices of companion blocks).Answer: ΔT(t)=(t4+t+1)(t−2)4.M=diag(C(t4+t+1),C((t−2)2),C((t−2)2))orM=diag(C(t4+t+1),C((t−2)2),C(t−2),C(t−2)).That is, C(t4+t+1)=010000100001−1−100,C((t−2)2)=[01−44],C(t−2)=[2].
Problem 5.3
Let V be a vector space of dimension 9 over Q, and let T:V→V have minimal polynomialmT(t)=(t2+t+1)3(t+1).(a) Determine the characteristic polynomial ΔT(t).(b) Determine the rational canonical form M (as a block diagonal matrix of companion blocks).Answer: ΔT(t)=(t2+t+1)4(t+1),M=diag(C((t2+t+1)3),C(t2+t+1),C(t+1)).That is, (t2+t+1)3=t6+3t5+6t4+7t3+6t2+3t+1,C((t2+t+1)3)=010000001000000100000010000001−1−3−6−7−6−3,C(t2+t+1)=[01−1−1],C(t+1)=[−1].