Parametrizing the value of the vector minimax with dependent constraints (Q1569251)

The authors deal with the concept of the value of multicriteria minimax problems with dependent variables. Three variants of its formal definition as a set of the best guaranteed estimates are given. A case is described in which all these variants yield the same result; a method for parametrizing the corresponding Slater set by the values of the minimax of the inverse logical convolution of partial criteria is suggested. It is shown that the minimizing strategies are also determined by realizations of the minimax of the linear convolution.