针对双响应优化设计中如何权衡每个响应的权重缺乏理论依据、没有考虑优化模型结果的稳健性和最优性的难题,提出了基于改进的满意度函数法双响应曲面稳健设计新方法。通过引入熵权理论对传统双响应曲面主要凭经验确定权重的方法进行了改进,构建了改进的综合满意度函数,并采用广义降阶梯度法优化对其求解,实现了将双响应优化问题转化为求综合满意度函数最大值的简化问题。实例表明,采用该方法能够得到全面稳健优化的设计方案,验证了该方法的有效性。
This paper aims to how to balance the weight of each response lack of theoretical basis and tough problem of model optimization during the process of dual response design optimization, it did not consider the results of robustness and optimal, proposes a new dual response surface method based on the improved satisfaction functional- based robust design. By introducing the entropy theory, it has made an improvement towards the traditional dual response method which determines weights by its experience, built up an improved overall satisfaction function and an optimized GRG(General Reduced Gradient) method for solving them, realized the simplifying of the issues of dual response optimization convert to pursuit the maximal value of overall satisfaction, the real example showed that this method can be fully optimized design robust which verified the effectiveness of the method.