The strongly compatible topology on \(\nu \)-generalized metric spaces (Q683623): Difference between revisions

Let \(X\) be a set and let \(\nu \geq 2\) be an integer number. A symmetric function \(d \colon X \times X \to [0, \infty)\) is said to be a \(\nu\)-generalized metric if \[ 0 = d(x, x) < d(x, y) \leq d(x, a_1) + d(a_1, a_2) + \ldots + d(a_\nu, y) \] holds for all distinct \(x\), \(y\), \(a_1\), \(\ldots\), \(a_{\nu} \in X\). A topology \(\tau\) on \(X\) is strongly compatible with \(d\) if the following statements are equivalent for every net \(\{x_\alpha\} \subset X\) and every point \(x \in X\): {\parindent=0.7cm \begin{itemize}\item[(1)] \(\{x_{\alpha}\}\) converges to \(x\) in \((X, \tau)\); \item[(2)] the equalities \[ 0 = \lim_{\alpha} d(x, x_{\alpha}) = \lim_{\alpha} \sup\{d(x_\alpha, x_\beta) \colon \beta \geq \alpha\} \] hold. \end{itemize}} Similarly \(\tau\) is compatible with \(d\) if (1) is equivalent to \(\lim_{\alpha} d(x, x_{\alpha}) = 0\). For every \(x \in X\), let \(\mathbf{A}_x\) be the set of all subsets \(A\) of \(X\) such that \(x \notin A\) and \(\operatorname{card} A = \nu + 2\). Define a function \(\delta \colon \mathbf{A}_x \to [0, \infty)\) by \[ \delta(A) := \max \{d(x, a_1) + \sum_{j=1}^{\nu + 1} d(a_j, a_{j+1})\}, \] where the maximum is taken over all numerations \(\{a_1, \ldots, a_{\nu + 2}\}\) of \(A \in \mathbf{A}_x\). Let \(\eta (x) := \inf \{\delta(A) \colon A \in \mathbf{A}_x\}\) and \[ \rho(x, y) := \begin{cases} 0 & \text{if } x = y\\ d(x,y) + \eta(x) + \eta(y) & \text{if } x \neq y \end{cases} \] for all \(x\), \(y \in X\). For every \(X\) with \(\operatorname{card} X \geq \nu + 3\) and every \(\nu\)-generalized metric \(d \colon X \times X \to [0, \infty)\), the author proves that \((X, \rho)\) is a metric space whose topology is strongly compatible with \(d\). It is interesting to note that there are \(\nu\)-generalized metrics which do not have any compatible topology, as was proved by the same author earlier. Moreover, it is shown that \((X, \rho)\) is complete if and only if \((X, d)\) is complete when the completeness of \((X,d)\) is understood in the sense that for every sequence \(\{x_n\} \subset X\) with \[ \lim_{n \to \infty} \sup_{m > n} d(x_n, x_m) = 0, \] there is \(x \in X\) satisfying the equality \(\lim_{n \to \infty} d(x, x_n) = 0\). Similar results are also obtained for the compactness of \((X, \rho)\) and \((X, d)\).