Some identities involving the Prouhet-Thue-Morse sequence and its relatives (Q519991)

From MaRDI portal
scientific article
Language Label Description Also known as
English
Some identities involving the Prouhet-Thue-Morse sequence and its relatives
scientific article

    Statements

    Some identities involving the Prouhet-Thue-Morse sequence and its relatives (English)
    0 references
    0 references
    0 references
    31 March 2017
    0 references
    Let \(s_k(n)\) be the sum of digits of the expansion of the integer \(n\) in base \(k\). The authors give some generalization of certain identities involving the classical Prouhet-Thue-Morse sequence \(t_n = (-1)^{s_2(n)}\) and identities given by H. Nieto and P. Bateman and D. Bradley. More precisely, they investigate the sequence of polynomials \(f^{\mathbf{u}}_{m,n}(x)=\sum_{i=0}^{k^n-1} \zeta_k^{s_k(i)}(x+\mathbf{u}(i))^m\), where the sequence \(\mathbf{u}\) satisfies the recurrence relation \(\mathbf{u}(ki+j)=P(\mathbf{u}(i))+jq\) for \(t=0,1,\dots,k-1\) and \(q\in V\), with \(V\) a finitely dimensional space over the field \(K\) and \(P: V\to V\) a linear endomorphism. An essential part of the paper (Section 2) is devoted to study of properties of the polynomials \(f^{\mathbf{u}}_{m,n}(x)\). For instance, how to reduce the computation of \(f^{\mathbf{u}}_{m,n}(x)\) via computation of its constant term, they give an explicit formula for this number, or how a recurrence relation satisfied by these constant terms leads to several curios identities involving the function \(s_2(n)\), etc. If \(N_k(i,n)\) denotes the number of occurrences of a digit \(i\) in the expansion of an integer \(n\) in the base \(k\) they define a related function \(N_{k,v}(n)=\sum_{i=1}^n N_k(v_i,n)\), where \(v=(v_1,\dots,v_b)\) is a \(b\)-tuple of pairwise distinct elements of \(\{1,\dots,2b-1\}\) with \(2b-1=k\). This function satisfies the identity \(\sum_{i=1}^{k^n-1} (-1)^{N_{k,v}(i)}i^m=0\) which resembles a basic identity for the Prouhet-Thue-Morse sequence in such a way that it leads to a partition of \(\{0,1,\dots,k^n-1\}\) into two disjoint sets \(P\) and \(Q\) with \(\sum_{i\in P}i^m=\sum_{i\in Q}i^m\) for each \(m=0,1,\dots,n-1\). The final section of the paper contains several conjectures generalizing the identities obtained in previous sections.
    0 references
    0 references
    0 references
    0 references
    0 references
    0 references
    Prouhet-Thue-Morse sequence
    0 references
    automatic sequence
    0 references
    sum of digits
    0 references
    identity
    0 references
    0 references
    0 references