置换矩阵
随机排列
组合数学
邻接表
编码(内存)
集合(抽象数据类型)
邻接矩阵
广义置换矩阵
数学
排列(音乐)
部分置换
离散数学
算法
计算机科学
位反转置换
基质(化学分析)
循环置换
特征向量
协方差矩阵
置换图
随机矩阵
理论计算机科学
矩阵代数
拓扑(电路)
作者
Ruolan CHENG,Tianxiang ZHAN,Yong Deng,Ignacio Javier Pérez
标识
DOI:10.1016/j.cja.2026.104569
摘要
Random Permutation Set Theory (RPST) generalizes Dempster-Shafer Theory (DST) by extending the power set event space to an ordered permutation event space. This extension naturally raises the question of how such ordered event structures should be encoded to represent and compute order-dependent uncertainty. Existing Random Permutation Set (RPS) encodings are mainly coordinate-based and do not directly capture the local structural relations between consecutive elements. To address this issue, this paper proposes an adjacency matrix encoding for RPS. Each permutation event is represented by the sum of a diagonal existence matrix and a directed order adjacency matrix, encoding support and immediate successor relations in a unified binary form. From a graph-theoretic perspective, nonempty events are reinterpreted as directed paths. Unique recoverability and structural and algebraic properties of the proposed encoding are investigated. Belief, plausibility, and commonality functions in RPS are revisited in the adjacency-matrix domain. Moreover, this encoding maps existing RPS operations into the same closed matrix space. The proposed encoding provides a new structural interpretation of RPST by linking the classical event encodings in probability theory and DST with the graph-based encoding of random graph sets. Finally, its practical applicability is demonstrated through a simulation-based aircraft threat assessment experiment.
科研通智能强力驱动
Strongly Powered by AbleSci AI