Bounds for total antieigenvalue of a normal operator英文文献资料.docVIP

BOUNDSFORTOTALANTIEIGENVALUE OFANORMALOPERATOR SK.M.HOSSEIN, K.C.DAS,L.DEBNATH, andK.PAUL Received 12January 2004 and inrevised form 22March 2004 Wegive analternative proof ofatheorem ofGustafson and Seddighin (1993) following the ideausedbyDasetal.inanearlier study ofantieigenvectors (1998). Theresult proved here holds forcertain classes ofnormal operators even ifthe space isin?nite dimensional. 2000 Mathematics Subject Classi?cation: 47B44, 47A63, 47B15. 1. Introduction. LetH beacomplex Hilbert space and letB(H) denote theBanach algebra ofallbounded linearoperators onH.Theconcept oftheangleofanoperator T was introduced by Gustafson [4, 5, 6] while studying perturbation theory of semi- groupgenerators. Fromthishasdevelopedwhatwecalloperatortrigonometry, whose theoryandapplicationsarestillevolving.Thepropertiesareintimatelyassociatedwith the numerical range W(T) of an operator T and the numerical range W(TT?) . Some relevantimportant resultscanbefoundin[3]. Thecosineofanoperator T inB(H) wasoriginally de?nedasfollows: cosT = inf Re(Tf,f) (1.1) Tff Tf=0 for arbitrary operators in a Banach semi-innerproduct space. Here we will restrict at- tentionprimarilytothecaseofT ∈B(H).Clearly,CosT isarealcosinede?nedforthe realpartofthenumerical rangeofT.Thetotalcosine


(Tf,f)
Tff |cos|T = inf (1.2) Tf=0 isalsode?ned. The expression (1.1) also denoted as the angle φ(T) measures the maximum (real) turninge?ectofT. ThequantityCosT hasanotherinterpretation asthe?rstantieigenvalue ofT ,where μ1(T)=Tinff=0ReT(Tff,ff). (1.3) The terminology “antieigenvalue” and “antieigenvector” was introduced byGustafson [7]intheyear1972. Krein [10] and Gustafson [7] have studied μ1(T) and indicated how the knowledge ofμ1(T)canbeusefulinthestudyofcertainintegraloperators, initialvalue problems  3878 SK. M. HOSSEIN ET AL. andsomeotherareas.Theupperboundforμ1(T)forT,a?nite-dimensional, strongly accretive (i.e.,ReT 0)normaloperatorwasob