Abstract: In order to get rid of the limitation of D-S combination rule and improve rules, a new combination rule based on the weighting factor of conjunctive and complementary pooling and Pignistic probability transformation is introduced. Beginning with the theory of the conjunctive and complementary pooling, the new formula of basic probability assignment is inferred. The evidence combination rule is obtained through the quantification research on the weighting factor of conjunctive and complementary pooling. With the purpose of reducing the influence of combination sequence and providing a reliable decision-making, a Pignistic probability transformation is used to reassign the basic probability assignments of multi-element propositions to each element. Through analyzing the number examples, and compared with other improved methods, the results show this new solution can efficiently solve the conflicting evidence, the loss of majority opinion, the robustness of algorithm, the fairness of distribution and the efficiency of decision.

Key words: D-S evidence theory, evidence combination rule, conjunctive and complementary pooling, Pignistic transformation, target recognition

摘要： 针对D-S证据组合公式及其改进公式的局限性，提出一种基于交补集权重和Pignistic概率变换的改进组合方法。基于交补集理论推导出新的基本概率分配函数，对交补集权重因子进行量化，得到基于交补集权重的证据组合公式，利用Pignistic概率变换法对已获得的各命题的信度值进行重新分配，以降低组合顺序对合成结果的影响，同时可获得更可靠的决策依据。实例分析结果表明，与其他改进方法相比，该组合方法在解决冲突证据、一票否决、鲁棒性、公平性和决策有效性等方面均有明显的优势。

关键词: D-S证据理论, 证据组合规则, 交补集, Pignistic变换, 目标识别

CLC Number: 

• N945