Partial inner product spaces. Theory and applications (Q1037527)

Partial Inner Product (PIP) spaces generalize and synthesize a lot of spaces appearing in functional analysis, such as rigged Hilbert spaces, scales of Hilbert or Banach spaces, etc. Their study is motivated by numerous applications in quantum physics or in signal processing, where one works with parametrized families of spaces and it is the whole system that counts rather than the individual spaces. The authors propose an axiomatic approach allowing a unified treatment of several questions involving indexed families of spaces and operators between them. One considers a complex vector space \(V\) and a linear compatibility relation on it, meaning a symmetric relation \(\#\) such that \(f\#g,\;f\#h \Rightarrow f\#(\alpha g+\beta h)\) for all \(f,g,h\in V\) and \(\alpha,\beta\in \mathbb C\). It follows that, for any \(S\subset V\), \(S^\#=\{g\in V : g\# f\;\forall f\in S\}\) is a linear subspace of \(V\), \(S\subset S^{\#\#}\) and \(S^{\#\#\#}=S^{\#}\). A key notion in the whole construction is that of an assaying subspace, meaning a subspace \(S\) of \(V\) such that \(S= S^{\#\#}\). It is shown that the family \(\mathcal F(V,\#)=\{V_r : r\in F\}\) of all assaying subspaces of \(V\), ordered by inclusion, is a complete involutive lattice with respect to the operations \(\bigwedge_{j\in J}V_j=\bigcap_{j\in J}V_j,\; \bigvee_{j\in J}V_j=\left(\sum_{j\in J}V_j\right)^\#\) and involution \(V_r\leftrightarrow V_{\bar r}=\big(V_r\big)^\# \) The involution is a lattice anti-isomorphism, i.e., \(\big(V_r\wedge V_s\big)^\# =V_{\bar r}\vee V_{\bar s} \) and \(\big(V_r\vee V_s\big)^\# =V_{\bar r}\wedge V_{\bar s} \). A partial inner product is a Hermitian form \(\langle\,\cdot\,|\,\cdot\,\rangle\) defined for all compatible pairs \(f\#g\) from \(V\). A PIP-space is a vector space \(V\) equipped with a linear compatibility relation and a partial inner product. A typical example is that of weighted \(\ell^p\) spaces. One considers on the space \(\omega\) of all complex sequences the linear compatibility \(x\# y\iff \sum_n|x_ny_n|<\infty\) and one defines the partial inner product by \(\langle x|y\rangle=\sum_n\bar{x}_ny_n\) for \(x=(x_n)\), \(y=(y_n)\) in \(\omega\) with \(x\# y\). For a sequence \(r=(r_n)\) of strictly positive numbers, one considers the space \(\ell^p(r)\) of all sequences such that \(\|x\|_r^{(p)}:=\left(\sum_n|r_n^{-1}x_n|p\right)^{1/p}<\infty\). In this case \(\ell^p(r)^\#=\ell^{\bar p}(\bar r),\) where \(1/p+1/\bar p=1\) and \(\bar r_n=r_n^{-1}\), \(n\in \mathbb N\). Similar examples can be constructed for spaces of locally integrable functions on \(\mathbb R\) or on \(\mathbb R^n\), including Köthe sequence spaces and Köthe function spaces. Other examples are the Bergman and Hardy spaces of analytic functions in the unit disk. Considering vector topologies on the spaces \(V_r\), these must agree with the algebraic duality. As a rule, the authors take the Mackey topology for a dual pair. The authors also study linear operators acting between PIP-spaces. A good idea of the topics treated in the book is given by the headings of the chapters: 1. General theory: Algebraic point of view; 2. General theory: Topological aspects; 3. Operators on PIP-spaces and indexed PIP-spaces; 4. Examples of indexed PIP-space; 5. Refinements of PIP-spaces; 6. Partial \(*\)-algebras and operators in a PIP-space; 7. Applications in mathematical physics (Dirac formalism, quantum mechanics, scattering, quantum field theory); 8. PIP-spaces and signal processing. Two appendices complete the contents: A. Galois connections and B. Some facts about locally convex spaces. Based on a long-term cooperation (starting in 1967) of one of the authors (JPA) with A.\,Grossmann (who furnished the authors with some unpublished notes, dubbed ``the skeleton'') and between the authors, the book will be of interest for researchers interested in function spaces as well as for those interested in applications in theoretical physics and signal processing.