Discrete optimization under interval uncertainty (Q1320714)

The proposed statement of the discrete optimization problem differs from the traditional one in the coefficients of the goal function -- instead of precise values closed intervals are used containing these coefficients. The principles of comparison and optimization of these interval values are based on the use of infinite-valued logic. Proved are: (a) three theorems, reducing the choice of an extremal one-interval value; (b) three theorems being used for the existence of a solution. All the presented examples of proposed solution techniques are connected to modifications of an assignment problem solution algorithm, using a branch-and-bound algorithm at the first step of the procedure. The article does not contain any numerical results.