一种基于能量函数的证据合成算法

Combination algorithm for evidence theory utilizing energy function

传统证据合成的计算量会随着证据个数急剧增加,限制了DS证据理论的广泛应用。从减少参与证据合成的焦元数量出发,提出了一种改进的DS证据理论的实用算法,以解决DS证据理论的实用化问题。该算法以焦元的能量函数、证据的平均能量函数作为选择抛弃焦元和剩余焦元的标准,同时将抛弃焦元的基本概率赋值重新分配给与之具有嵌套关系或相交关系的剩余焦元。该算法保留了抛弃焦元在证据合成过程中的生存权利,充分考虑了计算量和准确性。通过实例比较了该算法与其他几种近似算法的性能,验证了所提出算法的有效性和合理性。

The computational complexity of reasoning within the Dempster-Shafer（DS） theory of evidence is one of the major points of criticism in many practical applications.To solve such a problem,various approximation algorithms have been suggested.An improved practical algorithm is presented through reducing the number of focal elements involved.In this proposed algorithm,all focal elements of every piece of evidence are classified into dereliction and remainder,and the basic probability assignments of those derelictions are reassigned to the remainders when they are correlative or the dereliction is nested to the remainder.Furthermore,the effect of the dereliction is considered well to influence the combination in this paper,and the computation and accuracy are taken into account to develop this proposed algorithm.Finally,an illustrative example shows that the improved practical algorithm is effective and feasible by comparing with other approximations.