Strictly totally positive systems (Q1266089)

Let \(F(A)\) denote the set of all real-valued functions defined on \(A\subset \mathbb{R}\). A system \(U_n= \{u_0, \dots, u_n\} \subset F(A)\), is called a \(T\)-system if Cardinal \(A \geq n+1\), the functions in \(U_n\) are linearly independent on \(A\), and all the determinants of the square collocation matrices \[ M{u_0, \dots, u_n \choose t_0, \dots, t_n}: =\bigl(u_j(t_i); 0\leq i\leq m,\;0\leq j\leq n\bigr) \tag{1} \] with \(t_0< \cdots <t_n\) in \(A\), are positive. In case of all the determinants are nonnegative instead of positive, it is called a weak T-system or WT-system. If all the collocation matrices (1) are strictly totally positive, it is said to \(U_n\) is a strictly totally positive system or an STP-system. A system \(U_n\) is called a Markov system if \(U_k= (u_0, \dots, u_k)\) is a \(T\)-system for each \(k=0,1, \dots,n\). This paper is devoted to study the theory of T-systems and the linear spaces they generate. Algebraic methods are used to obtain splicing theorems for Markov and STP-systems. The following result, where \(A<B\) means that \(a<b\) for every \(a\in A\) and \(b\in B\), is related to Markov systems. Theorem: Let \(A,B \subset \mathbb{R}\) be such that \(A\setminus B< A\cap B<B \setminus A\) and Cardinal \(A\cap B\geq n\). If \(U_n \subset A\cup B\) which is a Markov system on \(A\) and also a Markov system on \(B\), then it is a Markov system on \(A\cup B\). The problem of extending the domain of definition of WT and T-systems, and their respectives spaces is studied in this article. Theorem: Let \(A,B,C\) be such that \(A<B<C\) and let \(U_n \subset F(B)\). If \(U_n\) is a Markov system (STP-system) on \(B\) and can be extended as a Markov system (STP-system) to \(A\cup B\) and as a Markov system to \(B\cup C\), then \(U_n\) can be extended as a Markov system (STP-system) to \(A\cup B\cup C\).