No label defined (Q4953944)

The Cartesian product of \(m\) (\(m>0\)) open intervals \((x_{i},y_{i})\), \(x_{i},y_{i} \in {\mathbb {R}}\) (\(i=1,2,\dots ,m.\)), is called finite \(m\)-dimensional open interval, and its closure -- finite \(m\)-dimensional closed interval. Let \(D\subset {\mathbb R}^{m}\) be an open interval and let the function \(p:D\rightarrow {\mathbb R}\) be continuous on \(D\). The function \[ Z[p]:(D\times {\mathbb R})\setminus \{(c,\alpha)\in D\times {\mathbb R}:p(c)= \alpha\}\rightarrow {\mathbb R}^{m} \] is called a zone function of \(p\) if its values for all pairs \((c,\alpha)\) from its domain are open intervals and \[ (c \in Z[p](c,\alpha) \wedge x \in Z[p](c,\alpha)\cap D) \Rightarrow (p(x) \neq \alpha). \] The author gives a survey of results in interval analysis during the decade of the 1990's presenting the main applications of the theory, namely numerical methods and their algorithms: (i) to find one of the solutions of a nonlinear system of \(n\) equations with \(m\) unknowns in a given \(m\)-dimensional interval, (ii) to find (i.e. to approximate) the global (in a closed \(m\)-dimensional interval) minimum of a function, value of which at each point of the interval is defined as maximum of correspondent values of \(m\) given functions, and (iii) to find a feasible approximating solution of the problem of nonlinear programming defined in an \(m\)-dimensional closed interval. When all functions appearing in such problems may be built from the well-known univariate real elementary functions, the interval algorithms solving these problems may be realized using the Maple or Mathematica programming languages, because they can handle the zone functions.