Nash交渉ゲームモデルによる設計者の志向にそった制御系の最適化

抄録

This paper explains an optimum design concept for a designer's intentions in controller design. We consider multi-objective optimization through the Nash bargaining model ; an optimum solution attainable from Nash's study. We then optimize a feedback gain of LQR(Linear Quadric Regulator)on a motor-mass system to show the concept's effectiveness. In achieving the LQR optimization, we aim to simultaneously reduce output errors and input amplitude. Based on the Nash bargaining model, the criterions, output errors and input amplitude, are represented by numeric values, called utilities, which formulate a compromise point. Afterwards, a design solution is evaluated from this point. The resultant solution satisfies a designer's intention.