Skip to content

2024春 高代下 final(H) (回忆版)

Problem 1

Let X be the set of all symmetric n×n matrices over R , and YX the subset of all positive semidefinite matrices

1-(i)

For two matrices A,BX , define AB if ABY. Prove that is a partial order, ie

  • (1) AA
  • (2) AB and BA implies A=B
  • (3) AB and BC implies AC

1-(ii)

Prove that Y is convex, ie. for A,BY , we have tA+(1t)BY for all real t[0,1]

1-(iii)

Let A,BY. Suppose that AB=BA. Prove that ABY

1-(iv)

Let

Z:={AX:tr(AB)0 for all BY},

where tr means trace. It is a known fact that tr(AB)=tr(BA) for matrices A, B with compatible sizes. Prove that Z=Y

Problem 2

2-(i)

Let X:=(0.01,2π)×(0.01,2π). Consider f:XR given by f(x,y)=(cosx)(cosy). Find all its local maximals(both points and values) and local minimals(both points and values)

2-(ii)

Let X be the set of real 2×2 orthogonal matrices. Consider f:XR given by AtrA2 where tr means trace. Find all its go global maximum and minimum(both points and values)

Problem 3

We define an equivalence relation on Mn(k) by

AB if B=SAS for some invertible SMn(k)

3-(i)

For k=C and n=2 , find a complete set of representatives for Mn(k)/

3-(ii)

For k=R and n=2 , find a complete set of representatives for Mn(k)/

Problem 4

Let M be a finite inner product space over C. For an ordered basis B , let GB be its Gram matrix. A pair of dual bases consists of an ordered basis B=(b1,,bm) of M and another ordered basis C=(c1,,cm) of M such that

bi,CjM=δij

where δij:=1 if i=j and δij:=0 if ij

4-(i)

Let b1,,bm and c1,,cm be dual bases. Give a simple formula for an element vM as a linear combination of b1,,bm

4-(ii)

Let b1,,bm be an ordered basis of M. Prove that there exist c1,,cmM such that b1,,bm and c1,,cm are dual bases

4-(iii)

Prove that, if B and C are dual bases, then GB and Gc are inverses to each other

4-(iv)

Disprove that, if GB and GC are inverses to each other, then B and C are dual bases

Problem 5

Consider square matrices over C. It is a known fact that every matrix is upper triangularizable. Prove the following statements

5-(i)

Every matrix is unitarily triangularizable

5-(ii)

Eigenspaces with distinct eigenvalues of normal matrix are orthogonal

5-(iii)

Every normal matrix is unitarily diagonalizable

5-(iv)

A matrix is unitarily diagonalizable if and only if it is normal

Problem 6

Let k:=c. For a AMn(k), let

eA:=i=0Aii!,

where we adopt the convention that A=I. It is known facts that eA always exists (the series converges) and eA+B=eA+eB for [A,B]=0. A logarithm of A is a matrix B such that eB=A

6-(i)

Find all logarithms of the 2×2 identity matrix

6-(ii)

Let θR such that sinθ0. Let Aθ:=[cosθsinθsinθcosθ]

Find all logarithms of Aθ

6-(iii)

Prove that, for a unitary matrix A, there exists a skew -Hermitian (Bt=B) logarithm B

6-(iv)

Prove that a logarithm of A exist if and only if A is invertible

Problem 7

Let M be a finite k-module over an algebraically closed field k. Let g:=End(M). For each xg, let x=xs+xn be its (unique) Jordan-Chevalley decomposition in g. For each xg, let ad x:gg be the (a prior not necessarily linear) map given by y[x,y], where [x,y]:=xyyx

7-(i)

Prove that the map ad x is an element in End(g), ie. and x is a linear transformation

7-(ii)

Prove that the map ad xs is semisimple in End(g)

7-(iii)

Prove that the map ad xn is nilpotent in End(g)

7-(iv)

Prove that [adxs,adxn]=0 in End(g)

Problem 8

Let k:=C,d1, and C0,,Cd1k. Let p:=xd+Cd1xd1++c1x+C0 and suppose that p=(xλ1)d(xλt)dt for distinct λ1,,λtk. The companion matrix of p is

C:=[000C0100C1010C20001Cd1]d×d

8-(i)

For d=2, find an explicit similarity from C to its Jordan normal form

8-(ii)

Consider a recurrence f0=0,f1=1 and fn+2=afn+1+bfn, where a,b,fiC. Given an explicit formula (without using matrices) for fn in terms of a,b,n

8-(iii)

Prove that the only annihilating polynomial of C of degree at most d1 is 0

8-(iv)

Calculate the Jordan normal from form of C