Recent progress in polyhedral harmonics (Q1580457)

From MaRDI portal
scientific article
Language Label Description Also known as
English
Recent progress in polyhedral harmonics
scientific article

    Statements

    Recent progress in polyhedral harmonics (English)
    0 references
    0 references
    12 March 2001
    0 references
    Let \(\Omega\subset\mathbb{R}^n\) be a domain and \(\mu\) a probability measure with compact support \(K\subset\mathbb{R}^n\). A continuous function \(f:\Omega \to\mathbb{R}\) is said to have the \(\mu\)-mean value property if \(f\) satisfies, \[ f(x) :=\int_K f(x+ry)d \mu(y),\quad \forall x\in\Omega \text{ and }0<r<r_x, \tag{1} \] where \(r_x\) is a sufficiently small constant such that \((x+rK)\subset \Omega\) for each \(r\). Flatto and Friedman-Littman raised the problem of characterising the class of such functions \(f\), which will be denoted in the sequel by \(H_\mu(\Omega)\). The paper under review begins with some historical remarks, in which results given by Friedman-Littman are stated in Section 2. In the remaining Sections 3 to 7 the author provides a survey of works concerning the case where \(K\) is a polytope in \(\mathbb{R}^n\), focusing on the recent contributions to the subject given by himself. A closed convex polytope is a finite intersection of closed half-planes in \(\mathbb{R}^n\) which is bounded and has non-empty interior. By definition an \(n\)-dimensional polytope \(P\) is a finite union of closed convex polytopes. Note that such a set may be neither convex nor connected. For \(d=0,1,\dots,n\) let \(P(d)\) be the \(d\)-skeleton of \(P\), i.e., the union of all its faces of dimension \(<d\) and let, \[ m_d:= {1\over\bigl|P(d)\bigr|} \mu_d, \tag{2} \] where \(\mu_d\) is the Euclidean measure on \(P(d)\) with total mass denoted by \(|P(d)|\). Let \(H_{P(d)} (\Omega)\) be the class of all continuous functions in \(\Omega\) which satisfies the \(m_d\)-mean value property \((=P(d)\)-mean value property). The author discusses in Section 3 the following fundamental results due to himself: (i) The restriction map \(H_{P(d)} (\mathbb{R}^n)\to H_{P(d)} (\Omega)\) is an isomorphism and hence \(H_{P(d)}(\Omega)\) is independent of the domain of definition \(\Omega\). (ii) \(H_{P(d)} (\Omega)\) is a finite-dimensional linear space of polynomials. (iii) A basis of \(H_{P(d)} (\Omega)\) can be taken from homogeneous polynomials. (iv) Let \(\delta= ({d\over dx_1},\dots, {d\over dx_n})\) then \(H_{P(d)}(\Omega)\) admits a structure of \(R[\delta]\)-module, where \(R[\delta]\) is the ring of (p.d.e.) with constant coefficients. (v) If the complete symmetry group \(G\subset O(n)\) of \(P\) is irreductible, then \(H_{P(d)}(\Omega)\) is a finite-dimensional space of harmonics monomials. In Section 4 a characterization of the \(P(d)\)-mean value property in terms of (p.d.e.) is stated. Section 5 is devoted to the case of polytopes with symmetry. The results indicate a deep connection between the dimension of \(H_{P(d)}\) and the classification of regular polytopes given by Coxeter. The author investigates in section 6 the case of polytopes without assumptions on symmetry, starting with the case of mean value property over triangles. He claims that the case of a polygon or an \(n\)-dimensional simplex are to be considered next. The last Section 7 is concerned with four open problems on the structure of \(H_{P(d)}\) as an \(R(\delta)\)-module.
    0 references
    0 references
    0 references
    0 references
    0 references
    0 references
    polyhedral harmonics
    0 references
    mean value property
    0 references
    \(n\)-dimensional polytope
    0 references
    closed convex polytope
    0 references
    polynomials
    0 references
    polytopes with symmetry
    0 references