摘要
为提取和量化群体决策过程中群体偏好和一致性意见,合理控制和调整研讨过程,达到群体一致性最优收敛,提出一种基于模糊Petri网的研讨节点评价算法。算法基于模糊Petri网建立研讨模型,将研讨节点结构化分解成若干个证据和一个主张,用协商研讨树来描述节点间的关系和研讨过程,使用模糊Petri网来形式化表示研讨树中的不确定性知识。通过构建证据的可信度,用模态值来量化节点之间支持或反对的强度,定义FPN中并行计算的方式和证据可信度的传递规则,给出任意节点共识值的计算方法并依据共识值确定群体偏好和一致性意见。将算法应用于多个具体的研讨实例,实验结果表明,算法得到的共识值指标能准确反映群体的共识状态并与现实相符,从而较好地验证了算法的可靠性和有效性。
In the process of group decision-making,it is important to extract and quantify the group preference and consensus opinion, control and adjust the decision process,and get the optimal convergence of group consistency. Aiming at this problem,a novel evaluation algorithm for argumentative node based on fuzzy Petri Net is proposed. It is based on argumentative tree model in which structure of the argumentative node is decomposed into several pieces of evidence and a claim, and the deliberation dialogue tree is used to describe the relationships between nodes and processing of deduction. By using fuzzy Petri Net,the algorithm can present the uncertain knowledge in the deliberation dialogue tree formally ,and modal values is used to quantify the strength of support or opposition between nodes through building the reliability of the evidence. Meanwhile,it is defined that the method of parallel computing in FPN and the transfer rules of the evidence＇ s reliability. The value of consensus of any node can be obtained and used to determine group preference and consistency. Furthermore,experimental results and practical applieation eases show that the quality of the algorithm proposed is better than that of state- of-the-art methods and it can accurately reflect the consensus state of the group,and the reliability and validity of algorithm is also be proved well.
出处
《计算机技术与发展》
2017年第3期91-96,102,共7页
Computer Technology and Development
基金
广东省科技计划项目(2013B090200006)
关键词
协商研讨
模糊PETRI网
可信度
共识值
deliberation dialogue
fuzzy Petri Net
credibility
value of consensus
