Entropy numbers and eigenvalues of operators Radosław Szwedek Adam Mickiewicz University in Poznań Faculty of Mathematics and Computer Science Aleksander Pełczyński Memorial Conference 2014 The talk is based on a recent work with Mieczysław Mastyło
Spectral radius formula ◮ The spectrum σ ( T ) of T on a complex Banach space X � � σ ( T ) := λ ∈ C ; λ I X − T is not invertible in L ( X ) ◮ The essential spectrum σ ess ( T ) := σ ( T ) T is the coset of T in the Calkin algebra L ( X ) / K ( X ) . ◮ Gelfand’s spectral radius formula m →∞ � T m � 1 / m r ( T ) := sup | λ | = lim λ ∈ σ ( T ) ◮ The essential spectral radius m →∞ � T m � 1 / m r ess ( T ) := r ( T ) = lim ess
Eigenvalue sequence The Riesz part of the spectrum Λ( T ) is at most countable and consists of isolated eigenvalues of finite algebraic multiplicity. � � Λ( T ) := λ ∈ σ ( T ) ; | λ | > r ess ( T ) � � ∞ We assign an eigenvalue sequence λ n ( T ) n = 1 for T ∈ L ( X ) from � � the elements of the set Λ( T ) ∪ r ess ( T ) as follows: ◮ The eigenvalues are arranged in an order of non-increasing absolute values. ◮ Every eigenvalue λ ∈ Λ( T ) is counted according to its algebraic multiplicity. ◮ If T possesses less than n eigenvalues λ with | λ | > r ess ( T ) , we let λ n ( T ) = λ n + 1 ( T ) = . . . = r ess ( T )
Entropy numbers Definition The n -th entropy number ε n ( T ) of T ∈ L ( X , Y ) is defined by � � n � ε n ( T ) := inf ε > 0 ; T ( U X ) ⊂ { y i + ε U Y } , y i ∈ Y i = 1 ◮ Entropy numbers are monotonne 0 � . . . � ε 3 ( T ) � ε 2 ( T ) � ε 1 ( T ) = � T � ◮ The measure of non-compactness β ( T ) := lim n →∞ ε n ( T )
Carl-Triebel’s inequality (1980) � � ∞ Let λ n ( T ) n = 1 be an eigenvalue sequence of T ∈ L ( X ) on a complex Banach space X . ◮ Carl’s inequality √ | λ n ( T ) | � 2 e n ( T ) where e n ( T ) := ε 2 n − 1 ( T ) ◮ Carl-Triebel’s inequality � n � 1 / n � k ∈ N k 1 / ( 2 n ) ε k ( T ) | λ i ( T ) | � inf i = 1
Banach couple ◮ We call � A := ( A 0 , A 1 ) a Banach couple if both A 0 and A 1 are Banach spaces such that A 0 , A 1 ֒ → X For a given Banach couple � A , we define spaces ◮ intersection A 0 ∩ A 1 with the norm � � � a � A 0 ∩ A 1 = max � a � A 0 , � a � A 1 ◮ sum A 0 + A 1 with the norm � � � a � A 0 + A 1 = inf � a 0 � A 0 + � a 1 � A 1 a = a 0 + a 1
Interpolation functor ◮ By T : � A → � B we denote an operator T : A 0 + A 1 → B 0 + B 1 , such that T | A j ∈ L ( A j , B j ) , j = 0 , 1 Definition By an interpolation functor we mean a mapping F : � B → B ◮ A 0 ∩ A 1 ⊂ F ( � A ) ⊂ A 0 + A 1 for any � A ∈ � B � � F ( � ⊂ F ( � A , � � B ∈ � B and T : � A → � ◮ T A ) B ) for any B
Interpolation functor of exponential type of θ For all interpolation functors F � � � T � F ( � B ) � C max � T � A 0 → B 0 , � T � A 1 → B 1 A ) →F ( � If in addition there exists θ ∈ ( 0 , 1 ) such that B ) � C � T � 1 − θ A 0 → B 0 � T � θ � T � F ( � A 1 → B 1 , A ) →F ( � then F is called of exponential type of θ . ◮ The real F ( · ) = ( · ) θ, q and complex F ( · ) = [ · ] θ interpolation functors are of exponential type of θ .
Recollect... ◮ The n -th entropy number ε n ( T ) of T ∈ L ( X , Y ) � � n � ε n ( T ) := inf ε > 0 ; T ( U X ) ⊂ { y i + ε U Y } , y i ∈ Y i = 1 ◮ The measure of non-compactness β ( T ) := lim n →∞ ε n ( T )
Interpolation of the measure of non-compactness β A delicate problem Let F be an interpolation functor of exponential type of θ . Does there exist a constant C > 0 such that for any T : � A → � B � � � � 1 − θ β � � θ ? T : F ( � A ) → F ( � β B ) � C β T : A 0 → B 0 T : A 1 → B 1 This question was answered positively ◮ for the real interpolation functor F ( · ) = ( · ) θ, q by Cobos, Fernández-Martínez and Martínez (1999), R.S. (2006), ◮ for the complex interpolation functor F ( · ) = [ · ] θ in the case where � B satisfies an approximation condition by Teixeira and Edmunds (1981), R.S. (2014).
Interpolation of entropy numbers fails A more delicate problem Let F be an interpolation functor of exponential type of θ . Does there exist a constant C > 0 such that for any T : � A → � B � � � C ε k 0 ( T : A 0 → B 0 ) 1 − θ ε k 1 ( T : A 1 → B 1 ) θ ? T : F ( � A ) → F ( � ε k 0 k 1 B ) ◮ This question was answered negatively for the real interpolation functor F ( · ) = ( · ) θ, q by Edmunds and Netrusov (2011). The reduction � B = � A
Recollect... ◮ The n -th entropy number ε n ( T ) of T ∈ L ( X ) � � n � ε n ( T ) := inf ε > 0 ; T ( U X ) ⊂ { x i + ε U X } , x i ∈ X i = 1 ◮ Carl’s inequality √ | λ n ( T ) | � 2 e n ( T ) where e n ( T ) := ε 2 n − 1 ( T ) ◮ Carl-Triebel’s inequality � n � 1 / n � k ∈ N k 1 / ( 2 n ) ε k ( T ) | λ i ( T ) | � inf i = 1
Interpolation variant of Carl-Triebel’s inequality Theorem (2013) Suppose that F is an interpolation functor of exponential type of θ . If T : � A → � A , then � �� � � � � � 2 e n ( T | A 0 ) 1 − θ e n ( T | A 1 ) θ � λ n T | F ( � A ) and � n � 1 / n � �� � � � � k 0 , k 1 ∈ N ( k 0 k 1 ) 1 / 2 n ε k 0 ( T | A 0 ) 1 − θ ε k 1 ( T | A 1 ) θ � λ i T | F ( � inf � � A ) i = 1
Generalizations of the spectral radius formula (1) ◮ Gelfand’s spectral radius formula m →∞ � T m � 1 / m | λ 1 ( T ) | = lim Definition Given T ∈ L ( X ) , the n -th approximation number is defined by � � a n ( T ) := inf � T − S � ; S ∈ L ( X ) , rank ( S ) < n ◮ König’s (1978) formula; a generalization for higher eigenvalues m →∞ a n ( T m ) 1 / m | λ n ( T ) | = lim
Generalizations of the spectral radius formula (2) ◮ The n -th entropy modulus g n ( T ) of T ∈ L ( X ) is given by � n � 1 / n � k ∈ N k 1 / ( 2 n ) ε k ( T ) =: g n ( T ) | λ i ( T ) | � inf i = 1 ◮ Makai-Zemánek’s formula (1982) � n � 1 / n � m →∞ g n ( T m ) 1 / m | λ i ( T ) | = lim i = 1 Problem In what form does it exist a formula for the spectral radius of T using the entropy numbers of powers of operators?
Main results - spectral entropy (1) Theorem (2013) � � Let X be a complex Banach space and T ∈ L ( X ) . If λ n ( T ) is an eigenvalue sequence of T , then � n � 1 / n � m →∞ ε k m ( T m ) 1 / m k − 1 / ( 2 n ) sup | λ i ( T ) | = lim n ∈ N i = 1 Definition We define the k -th spectral entropy number E k ( T ) by m →∞ ε k m ( T m ) 1 / m � ε k ( T ) E k ( T ) := lim
Main results - spectral entropy (2) Theorem (2013) m →∞ t 1 / m Fix t ∈ [ 1 , ∞ ) . If { t m } ⊂ N is such that lim = t , then m � n � 1 / n � m →∞ ε t m ( T m ) 1 / m t − 1 / ( 2 n ) sup | λ i ( T ) | = lim n ∈ N i = 1 Definition Define the spectral entropy map t �→ E t ( T ) of T as follows m →∞ ε t m ( T m ) 1 / m E t ( T ) := lim
Main results - spectral entropy (3) Proposition � � Let X be a complex Banach space and T ∈ L ( X ) . If λ n ( T ) is an eigenvalue sequence of T , then m →∞ ε m ( T m ) 1 / m = E 1 ( T ) = | λ 1 ( T ) | lim �� n � 1 / n i = 1 | λ i ( T ) | m →∞ e m ( T m ) 1 / m = E 2 ( T ) = sup lim √ 2 n ∈ N t →∞ E t ( T ) = r ess ( T ) lim
Entropy moduli Definition Let ϕ : [ 0 , ∞ ) → [ 0 , ∞ ) be a sub-multiplicative function. Given an operator T ∈ L ( X ) on a complex Banach space X , we define the entropy modulus g s ,ϕ ( T ) as follows k ∈ N k 1 / ( 2 s ) ϕ ( ε k ( T )) , g s ,ϕ ( T ) := inf s ∈ ( 0 , ∞ ) ◮ Denote by � ϕ the function on [ 0 , ∞ ) given by m →∞ ϕ ( u m ) 1 / m , ϕ ( u ) := lim � u � 0 ◮ � ϕ is sub-multiplicative and � ϕ � ϕ
Makai-Zemánek’s formula revisited Theorem (2013) Let X be an arbitrary complex Banach space and T ∈ L ( X ) . Assume that ϕ : [ 0 , ∞ ) → [ 0 , ∞ ) is a nondecreasing, sub-multiplicative and right-continuous function. Then t ∈ [ 1 , ∞ ) t 1 / ( 2 s ) � m →∞ g s ,ϕ ( T m ) 1 / m , inf ϕ ( E t ( T )) = lim s ∈ ( 0 , ∞ ) In particular, � n � 1 / n � t ∈ [ 1 , ∞ ) t 1 / ( 2 n ) E t ( T ) = inf | λ i ( T ) | i = 1
Interpolation of spectral entropy numbers holds Theorem (2013) If F be an interpolation functor of exponential type of θ , then for any T : � A → � A � � � E k 0 ( T : A 0 → A 0 ) 1 − θ E k 1 ( T : A 1 → A 1 ) θ T : F ( � A ) → F ( � E k 0 k 1 A )
Recommend
More recommend