一种改进的证据合成方法及其在模糊评估中的应用

Improved Evidence Combination Method and its Application in Fuzzy Evaluation

针对证据理论在模糊评估中的应用问题,分析了经典D-S证据合成规则和现有改进方法的不足,提出了一种更为适宜的证据合成方法。该方法通过证据间的夹角余弦计算证据间的相似系数,从而获得证据的可信度。通过对证据权重和可信度的对比与融合,确定各证据在冲突概率分配中的权重。通过冲突概率的全局分配对证据进行合成。实例分析表明,该方法不仅能够处理高度冲突证据,还能保证指标间的相对重要性不被可信度所过度地影响,从而提高了模糊评估的可靠性与合理性,降低决策风险。

To apply the Evidence Theory to Fuzzy Evaluation,an improved evidence combination method is proposed by analyzing the shortcomings of the classical D-S evidence combination rule andthe existing improved methods. Firstly,the credibility of the evidence is obtained by calculating thesimilar coefficient of the Evidences. Then,the distribution coefficients of the conflict probability are determined by comprehensive consideration of the weight and credibility of the evidence. At last,the evidence combination is carried through the overall distribution of the conflict probability. Example analysis shows that the proposed method can not only deal with the high conflict evidences,but alsocan enhance the reliability and rationality of the evaluation by preventing the dominance of the keyindex from being changed by the credibility.