Question and Answers Forum

All Questions      Topic List

Algebra Questions

Previous in All Question      Next in All Question      

Previous in Algebra      Next in Algebra      

Question Number 201343 by gabi last updated on 05/Dec/23

Answered by aleks041103 last updated on 05/Dec/23

A^3 +A^2 +A=0  λ is eigenvalue  ⇒λ^3 +λ^2 +λ=λ(λ^2 +λ+1)=0  ⇒λ=0,w,w^∗   where w=e^(2πi/3)  and w^∗  is the complex  conjugate of w.  let A=PLP^( −1) , where L is the Jordan normal  form of A.  ⇒P(L^3 +L^2 +L)P^( −1) =0  ⇒L^3 +L^2 +L=0  where L is block−diagonal with Jordan blocks  i.e.  L=diag(J_(11) ,..,J_(1a) ,J_(21) ,...,J_(2b) ,J_(31) ,...,J_(3c) )  where J_(ks)  is a Jordan block with k−th  eigenvalue  ⇒J_(ks) ^( 3) +J_(ks) ^( 2) +J_(ks) =0  J_(ks) =λ_k E_s +S_s   where E_s  is the identity over the s×s square  matrices and S_s  is the s×s square matrix  with 0s everywhere, except on the top  off−diagonal.  ex.:  J_(14) = [(λ_1 ,1,0,0),(0,λ_1 ,1,0),(0,0,λ_1 ,1),(0,0,0,λ_1 ) ]  J_(14) =λ_1  [(1,0,0,0),(0,1,0,0),(0,0,1,0),(0,0,0,1) ]+ [(0,1,0,0),(0,0,1,0),(0,0,0,1),(0,0,0,0) ]=  =λ_1 E_4 +S_4     It is good to note that  S_s ^k =0, if k≥s  and {S_s ^k }_(k=0) ^(s−1) ={E_s }∪{S_s ^k }_(k=1) ^(s−1)  is lin.indep.    Then  J_(ks) ^( 3) +J_(ks) ^( 2) +J_(ks) =0  ⇒(λE+S)^3 +(λE+S)^2 +λE+S=0  λ^3 E+3λ^2 S+3λS^2 +S^3 +λ^2 E+2λS+S^2 +λE+S=0  ⇒(λ^3 +λ^2 +λ)E+(3λ^2 +2λ+1)S+(3λ+1)S^2 +S^3 =0  (E,S∈M_(s×s) )  if s>3:  {E,S,S^2 ,S^3 }⊆{S_s ^k }_(k=0) ^(s−1)  is lin.ind.  ⇒ { ((λ^3 +λ^2 +λ=0)),((3λ^2 +2λ+1=0)),((3λ+1=0)),((1=0)) :} which is imposs.  if s=3:  S^3 =0, {E,S,S^2 } are lin.indep.  ⇒ { ((λ^3 +λ^2 +λ=0)),((3λ^2 +2λ+1=0)),((3λ+1=0)) :} which is imposs.  if s=2:  ⇒S^3 =S^2 =0 and {E,S} are lin.ind.  ⇒ { ((λ^3 +λ^2 +λ=0)),((3λ^2 +2λ+1=0)) :}         1)⇒λ=0,w,w^∗  ⇒ λ^2 =−λ−1 or λ=0         replace in 2)⇒ −3(λ+1)+2λ+1=−λ−2⇒λ=−2         but (−2)^3 +(−2)^2 +(−2)≠0         or 3(0)^2 +2(0)+1≠0  ⇒imposs.    ⇒s=1 ⇒ the jordan blocks are numbers  ⇒A is diagonalisable.    ⇒rg(A)=dim(ker(A−wE))+dim(ker(A−w^∗ E))  A is diagon.  ⇒ker(A−wE)=span({v_1 ,...,v_k })  where {v_1 ,...,v_k } are lin.ind. and   Av_m =wv_m   ⇒(Av_m )^∗ =A^∗ v_m ^∗ =Av_m ^∗ =(wv_m )^∗ =w^∗ v_m ^∗   where it is used that A^∗ =A, because  A∈M_(n×n) (R).  ⇒{v_1 ^∗ ,...,v_k ^∗ } are eigenvectors corresponding  to eigenvalue w^∗ ≠w.  ⇒span({v_1 ^∗ ,...,v_k ^∗ })≤ker(A−w^∗ E)  let u∈ker(A−w^∗ E)/span({v_1 ^∗ ,...,v_k ^∗ }) and  u≠0  then, Au=w^∗ u  ⇒Au^∗ =wu^∗ ⇒u^∗ ∈span({v_1 ,...,v_k })  ⇒u∈span({v_1 ^∗ ,...,v_k ^∗ })  ⇒u∈[span({v_1 ^∗ ,...,v_k ^∗ })]∩[ker(A−w^∗ E)/span({v_1 ^∗ ,...,v_k ^∗ })]={0}  ⇒u=0  ⇒ contradiction    ⇒ker(A−w^∗ E)=span({v_1 ^∗ ,...,v_k ^∗ })    ⇒rg(A)=dim(ker(A−wE))+dim(ker(A−w^∗ E))=  =dim(span({v_1 ,...,v_k }))+dim(span({v_1 ^∗ ,...,v_k ^∗ }))=  =k+k=2k  ⇒2∣rg(A)

A3+A2+A=0λiseigenvalueλ3+λ2+λ=λ(λ2+λ+1)=0λ=0,w,wwherew=e2πi/3andwisthecomplexconjugateofw.letA=PLP1,whereListheJordannormalformofA.P(L3+L2+L)P1=0L3+L2+L=0whereLisblockdiagonalwithJordanblocksi.e.L=diag(J11,..,J1a,J21,...,J2b,J31,...,J3c)whereJksisaJordanblockwithktheigenvalueJks3+Jks2+Jks=0Jks=λkEs+SswhereEsistheidentityoverthes×ssquarematricesandSsisthes×ssquarematrixwith0severywhere,exceptonthetopoffdiagonal.ex.:J14=[λ11000λ11000λ11000λ1]J14=λ1[1000010000100001]+[0100001000010000]==λ1E4+S4ItisgoodtonotethatSsk=0,ifksand{Ssk}k=0s1={Es}{Ssk}k=1s1islin.indep.ThenJks3+Jks2+Jks=0(λE+S)3+(λE+S)2+λE+S=0λ3E+3λ2S+3λS2+S3+λ2E+2λS+S2+λE+S=0(λ3+λ2+λ)E+(3λ2+2λ+1)S+(3λ+1)S2+S3=0(E,SMs×s)ifs>3:{E,S,S2,S3}{Ssk}k=0s1islin.ind.{λ3+λ2+λ=03λ2+2λ+1=03λ+1=01=0whichisimposs.ifs=3:S3=0,{E,S,S2}arelin.indep.{λ3+λ2+λ=03λ2+2λ+1=03λ+1=0whichisimposs.ifs=2:S3=S2=0and{E,S}arelin.ind.{λ3+λ2+λ=03λ2+2λ+1=01)λ=0,w,wλ2=λ1orλ=0replacein2)3(λ+1)+2λ+1=λ2λ=2but(2)3+(2)2+(2)0or3(0)2+2(0)+10imposs.s=1thejordanblocksarenumbersAisdiagonalisable.rg(A)=dim(ker(AwE))+dim(ker(AwE))Aisdiagon.ker(AwE)=span({v1,...,vk})where{v1,...,vk}arelin.ind.andAvm=wvm(Avm)=Avm=Avm=(wvm)=wvmwhereitisusedthatA=A,becauseAMn×n(R).{v1,...,vk}areeigenvectorscorrespondingtoeigenvalueww.span({v1,...,vk})ker(AwE)letuker(AwE)/span({v1,...,vk})andu0then,Au=wuAu=wuuspan({v1,...,vk})uspan({v1,...,vk})u[span({v1,...,vk})][ker(AwE)/span({v1,...,vk})]={0}u=0contradictionker(AwE)=span({v1,...,vk})rg(A)=dim(ker(AwE))+dim(ker(AwE))==dim(span({v1,...,vk}))+dim(span({v1,...,vk}))==k+k=2k2rg(A)

Commented by aleks041103 last updated on 05/Dec/23

We could′ve solved it in an even easier manner  let the minimal polynomial of A be m_A (x).  (by def m_A (x) is the lowest order polynomial  that has the property m_A (A)=0)  since f(A)=A^3 +A^2 +A=0, then m_A ∣f.  the spectrum of f(x)=0 is x=0,w,w^∗ .  ⇒f has a simple spectrum.  since m_A ∣f⇒m_A  has a simple spectrum too.  There is a theorem, which we essentially  proved for this specific case, that states  that the matrix A is diagonalizable  if and only if the minimal polynomial  of A has simple spectrum.  Thus, since m_A  has simple spectrum  then A is diahlgonalisable.  From here on forward it is the same.

WecouldvesolveditinaneveneasiermannerlettheminimalpolynomialofAbemA(x).(bydefmA(x)isthelowestorderpolynomialthathasthepropertymA(A)=0)sincef(A)=A3+A2+A=0,thenmAf.thespectrumoff(x)=0isx=0,w,w.fhasasimplespectrum.sincemAfmAhasasimplespectrumtoo.Thereisatheorem,whichweessentiallyprovedforthisspecificcase,thatstatesthatthematrixAisdiagonalizableifandonlyiftheminimalpolynomialofAhassimplespectrum.Thus,sincemAhassimplespectrumthenAisdiahlgonalisable.Fromhereonforwarditisthesame.

Terms of Service

Privacy Policy

Contact: info@tinkutara.com