Lefschetz numbers and geometry of operators in \(W^*\)-modules (Q1364870)

From MaRDI portal
scientific article
Language Label Description Also known as
English
Lefschetz numbers and geometry of operators in \(W^*\)-modules
scientific article

    Statements

    Lefschetz numbers and geometry of operators in \(W^*\)-modules (English)
    0 references
    4 November 1997
    0 references
    The aim of this cryptically written article is to define cyclic-homology-valued Lefschetz `numbers' for a (unitary) endomorphism on an \(A\)-elliptic complex \((E,d)\), \(A\) a \(W^*\)-algebra, the chain \(A\)-modules expressed as \({\mathcal{E}}_{k}\). No proofs are given. Earlier papers by the second-named author were for \(A\) a \(C^*\)-algebra, \(E\) a complex on a \(G\)-invariant manifold \(X\), \(G\) a compact group, and the endomorphism a representation of \(G\); a \(K_{0}(A) {\otimes} {\mathbb C}\)-valued Lefschetz number \(L_{1}\) was defined by the character-valued Atiyah-Singer index \(\text{ind}^{X}_{A,G}\). The authors consider a unitary operator \(U = \int_{S^{1}} {e^{i \phi}dP(\phi)}\) on a projective finitely generated Hilbert \(A\)-module \(\mathcal P\) (a Fredholm module over \(A\) since \((E,d)\) is elliptic), \(P\) being a measure taking values in the projections of \(A\). The definition given for \(L_1\) does not make sense and the definition in earlier papers by the second-named author is not applicable here. Taking a finite sum approximation of \(U\) by \(\sum_k { e^{i\phi_{k}}P(E_{k})}\), it is easily checked taking into account that \(P^{2} = P\), that \(\sum_{k}{P(E_{k}){\otimes}\dots {\otimes}P(E_k}) e^{i \phi_{k}}\), where there are \(2l\) copies of \(P(E_k)\), is in \(HC_{2l}(M(n,A))\). The authors define \(L_{2l}({\mathcal{E}},U)(\phi)\) as the trace of \(\int_{S^{1}}{{P(E_{k}) {\otimes}\dots {\otimes}P(E_k}) e^{i\phi_{k}}d\phi}\) so that \(L_{2l}({\mathcal{E}},U) \in HC_{2l}(A)\). The authors claim in their theorem that \(L_{2l}({\mathcal{E}},U) =\int_{S^1} {ch^{o}_{\ast} L_{1}({\mathcal{E}},U)(\phi) d\phi}\) where \(ch\) denotes a Chern character from \(K^0 \to HC_{2l}\), so supposedly \(L_{1}\) maps \(S^1\) to \(K_{0}(A)\). The article needs considerable revision. The authors also address some of the difficulties occurring when the modules are not self-dual, and for \(C\)*-modules without the group assumption.
    0 references
    cyclic-homology-valued Lefschetz numbers
    0 references
    \(W^*\)-module
    0 references
    elliptic complex
    0 references
    \(K\)-theory
    0 references
    Lefschetz number
    0 references
    \(C^*\)-module
    0 references
    \(A\)-elliptic complex
    0 references
    character-valued Atiyah-Singer index
    0 references
    0 references
    0 references

    Identifiers

    0 references
    0 references
    0 references
    0 references
    0 references
    0 references
    0 references
    0 references
    0 references
    0 references
    0 references