Overlaps in free partially commutative monoids (Q2639969)

From MaRDI portal
scientific article
Language Label Description Also known as
English
Overlaps in free partially commutative monoids
scientific article

    Statements

    Overlaps in free partially commutative monoids (English)
    0 references
    0 references
    0 references
    1991
    0 references
    A free partially commutative monoid M(\(\Sigma\),\(\Theta\)) is given through a finite alphabet \(\Sigma\) and an independence relation \(\Theta\subseteq \Sigma \times \Sigma\), which is symmetric and irreflexive. Then M(\(\Sigma\),\(\Theta\)) is the quotient monoid \(\Sigma^*/\equiv\), where \(\equiv\) denotes the Thue congruence induced by the commutation rules \(T_{\Theta}=\{(ab,ba)|\) (a,b)\(\in \Theta \}\). The subject of this paper is the set of overlaps of elements in free partially commutative monoids, and the focus is on how properties of overlaps in free monoids generalize to this situation. For a word \(u\in \Sigma^*\) let alph(u) denote the set of letters that occur in u. The trace [u] is a connected element of M(\(\Sigma\),\(\Theta\)) if alph(u) cannot be decomposed as \(alph(u)=A_ 1\cup A_ 2\), \(A_ 1\neq \emptyset \neq A_ 2\), such that \(A_ 1\times A_ 2\subseteq \Theta\). For \(x\in \Sigma^*\), \(OVL([x])=\{[u]|\) \(x\equiv uv_ 1\equiv v_ 2u\) for some \(v_ 1,v_ 2\neq e\}\) is the set of congruence classes of overlaps of [x] in the monoid M(\(\Sigma\),\(\Theta\)). The following results are obtained. For each element [x], the set of congruence classes OVL([x])\(\cup \{[x]\}\) is a lattice under the prefix-ordering \(\leq\) in M(\(\Sigma\),\(\Theta\)). If [x] is a connected element, then OVL([x]) is already a lattice under \(\leq\) (Theorem 3.3). If [x] is not connected, then the lattice OVL([x])\(\cup \{[x]\}\) is the direct product of the lattices \(OVL([x_ i])\), each with a new ``top'' element \([x_ i]\) adjoined, where the \([x_ i]\) are the connected components of [x]. Finally, an algorithm is presented that, given a connected element [x] as input, determines a maximum overlap of [x] in time linear in the length of x (Theorem 4.3).
    0 references
    0 references
    0 references
    0 references
    0 references
    0 references
    0 references
    0 references
    finite alphabet
    0 references
    independence relation
    0 references
    Thue congruence
    0 references
    commutation rules
    0 references
    overlaps of elements
    0 references
    free partially commutative monoids
    0 references
    overlaps in free monoids
    0 references
    connected element
    0 references
    congruence classes
    0 references
    prefix-ordering
    0 references
    direct product
    0 references
    lattices
    0 references
    connected components
    0 references
    algorithm
    0 references
    maximum overlap
    0 references
    time linear
    0 references
    length
    0 references
    0 references