Existence of maximizable quasiconcave functions on paracompact convex spaces (Q5966472): Difference between revisions

Let X be a nonempty, paracompact convex space. Suppose that (i) for each \(x\in X\), Tx is a nonempty convex subset of usc quasiconcave functions on X; (ii) for each usc quasiconcave function f on X, \(T^{-1}f\) is open in x; and (iii) there exists a nonempty, convex, compact set \(X_ 0\subset X\), and a nonempty, compact set \(K\subset X\) such that for each \(x\in X\setminus K\) and \(f\in Tx\), \(f(x)<\max \{f(y):y\in X_ 0\}.\) The main result of this paper asserts that there exist \(x\in K\) and \(f\in Tx\) such that f attains its maximum at x. This generalizes a theorem of Fan (who considers concave functions) and a theorem of Simons (who considers compact X). An example is included in showing that this is a strict generalization.