Abstract
There exists at least one pure strategy for each player of zero-sum two-person matrix-game which, if used against anyone of his opponent's optimal mixed strategic, yields the value of the game. This assertion will be extend to games where the payoff function is quasi concave quasi convex and upper/lower semi continuous. The result will be applied to games where one of the two players takes maximum advantage of the mistakes, if any, of his opponent. It will be shown that this problem is equivalence to a problem of Lexicographic multi objective Programming.

This publication has 8 references indexed in Scilit: