Abstract
Advanced materials for the semiconductor industry are being developed every month. To reduce the development period, we constructed a novel material design method based on linear programming, combinatorial optimization, and graph theory. This method (called weak conditioned combinatorial linear programming) enabled us to find composition candidates that satisfy a number of properties at the same time. By defining the solution area as a function of the combination index, the optimum formulations were acquired. This optimization could be done by newly developed user-friendly software. This system is applicable to optimization of materials with complex properties and time-series properties such as creep. The method enabled us to develop materials satisfying target values efficiently.
