Best approximation and best proximity pair theorems (Q2433602)

Let \(E\) be a normed space. For two subsets \(A,\) \(B\) of \(E\) let \(d\left( A,B\right) :=\inf \) \(\{\left\| x-y\right\| :x\in A\) and \(y\in B\}.\) If \(T\) is a mapping with domain \(A\) and values in \(E,\) a best approximation theorem provides sufficient conditions for the existence of an element \( x_{0}\in A\) (named best approximant) such that \(d(x_{0}\) \( ,Tx_{_{0}})=d(Tx_{0},A)\). A best proximity pair theorem provides the conditions under which for a mapping \(T:A\rightarrow B\) there exists \( x_{0}\in A\) such that \(d(x_{0},Tx_{0})=d(A,B)\). The author applies a proximity pair theorem published in [Acta Sci. Math. (Szeged) 63, 289--300 (1977; Zbl 0909.47042)] to the study of the existence of approximate equilibrium point in game theory.