Optimization of system reliability using 0-1 linear programming with interval coefficients

The authors propose a method for solving 0-1 linear programming problems with both generalized upper bounding (GUB) structures and interval coefficients without transforming them into crisp problems, keeping coefficients as intervals on the process of the calculation. In this method, they define the degree of satisfying the inequality relationship between intervals with minimal loss of fuzzy information. The proposed method is utilized for optimization of system reliability with fuzzy state to present an approximate solution and the interval objective function value. To demonstrate the effectiveness, a numerical example is shown and compared with conventional methods. [In Japanese.]