Pareto-optimal security strategies as minimax strategies of a standard matrix game (Q1306667)

In a multicriteria matrix game the entries in the payoff matrix are vectors, as opposed to real numbers in a standard matrix game. \textit{D. Ghose} and \textit{U. R. Prasad} [J. Optim. Theory Appl. 63, 167-189 (1989; Zbl 0662.90093)], introduced Pareto-optimal strategies in multicriteria matrix games, and \textit{D. Ghose} [ibid. 68, 463-481 (1991; Zbl 0697.90088)], characterized the Pareto-optimal security strategies of player 1 as minimax strategies in a weighted zero-sum game with a single criterion. This weighted zero-sum game is not yet a standard matrix game in that player 2 still selects a tuple of mixed strategies, one for each of the separate criteria. In the paper under review the author takes the final additional step toward standard matrix games, reducing the problem of finding a Pareto-optimal security strategy to finding a minimax strategy in a matrix game. It is perhaps significant to mention that the strategy space of player 2 grows exponentially with the data input: if player 2 has \(n\) pure strategies in the \(k\)-criterion game, he has \(n^k\) pure strategies in the resulting weighted matrix game.