On Farkas' lemma and related propositions in BISH (Q2067640)

This paper works within the framework of constructive mathematics, in the tradition of \textit{E. A. Bishop} [Foundations of constructive analysis. Maidenhead, Berksh.: McGraw-Hill Publishing Company, Ltd. (1967; Zbl 0183.01503)]; the paper analyzes the constructive validity of Farkas' lemma. \textit{J. Farkas}' lemma [J. Reine Angew. Math. 124, 1--27 (1901; JFM 32.0169.02)] states that for any real \(m\times n\)-matrix \(A\) and any point \(b\in R^{m}\), exactly one of the following statements is true. \begin{itemize} \item[i)] \(\exists\xi\in R^{m}(\xi\cdot A\ge0\wedge\xi\cdot b<0)\) \item[ii)] \(\exists q=(q_{1},\dots,q_{n})\in R^{n}(q_{i}\ge0(i=1,\dots,n)\wedge A\cdot q=b)\) \end{itemize} It is first shown that Farkas' lemma is constructively invalid. Then several constructively valid versions are obtained. These versions of Farkas' lemma are used to obtain constructive versions of several well-known classical results, including the Fredholm alternative for solvability of systems of linear equations, certain optimality criteria in linear programming, Stiemke's lemma, the superhedging duality, von Neumann's minimax theorem, and existence of solutions to two-person zero-sum games.