Stochastic matrices and a property of the infinite sequences of linear functionals (Q989063)

This paper deals with rectangular stochastic matrices. An \(n \times m\) nonnegative matrix such that the sum of the elements of each row is equal to one is called rectangular stochastic. In the first part of this work the author proves the following property of this class of matrices. Let \(T=(T_{ij})\) be an \(n \times m\) rectangular stochastic matrix such that for every pair of indices \((i,j)\), there exists an index \(l\) such that \(T_{il} \neq T_{jl}\). Then, there exists a real vector \(k=(k_1,k_2, \dots, k_n)^T\), \(k_i \neq k_j\), \(i \neq j\); \(0<k_i \leq 1\), such that \((T k)_i \neq (T k)_j\) if \(i \neq j\). In the second part of this paper, the author takes into account an infinite sequence of real functionals \(\{T_i\}_{i \in \mathbb{N}}\), \[ T_if = \int f(t)d \mu_t(i) =: G_f(i), \] corresponding to a sequence of probability measures \(\{\mu_{(\cdot)}(i) \}_{i \in \mathbb{N}}\), on the Borel \(\sigma\)-algebra \({\mathcal B}([0,1])\) such that \(\mu_{(\cdot)}(i) \neq \mu_{(\cdot)}(j)\), \(i,j \in \mathbb{N}\), \(i \neq j\). Then, by using the above property of rectangular stochastic matrices, the author proves constructively the existence of a real bounded one-to-one function \(f\) such that, for every \(i,j \in \mathbb{N}\), \(i \neq j\), \[ G_f(i):=\int f(t)d \mu_t(i) \neq \int f(t)d \mu_t(j)=:G_f(j). \] Finally, the author applies this result to the theory of positive operator valued measures which are used in quantum mechanics in order to generalize the concept of observable and which are a tool in quantum computation.