摘要
多专家决策算法是对多个专家的决策做出综合决策,使决策更符合实际情况。通常的算法使用各专家决策的加权求和法,由于各专家的权值在不同的情况下会有所不同,因此难以应用。本文提出采用基于支持向量机的非线性回归法进行综合决策,克服了权值变化的问题。多专家决策的另一途径是选取最佳的专家决策,本文采用基于支持向量机的非线性分类的方法,可以在不同情况下选取相应的最佳专家决策。本文提出的算法用于"隧道工程岩石力学参数综合处理系统",对若干子系统的结果进行综合,解决了专家权值变化的问题,取得了较精确的结果。
The multi-expert decision-making algorithm makes integrative decision, based on several experts' decisions, which is better than each individual decision. Usually the weighted sum strategy is adopted. But it is difficult to use in practice, because the weights assigned to each experts often vary in different conditions. In this paper, a nonlinear regression algorithm based on support vector machine is presented to make the integrative decision and overcome the variant weight problem. Another approach of multi-expert decision-making is finding the best expert's decision. In this paper, a nonlinear classfication method based on support vector machine is presented to provide best expert' s decision in various cases. The algorithm is applied to ' the integrated processing system for rock mechanical parameters in tunnel engineering', in which the integrative decision is obtained from several sub-systems, and good results are obtained.
出处
《模式识别与人工智能》
EI
CSCD
北大核心
2000年第2期165-169,共5页
Pattern Recognition and Artificial Intelligence
