基于图论的边缘计算信任评估优化模型

针对边缘计算中资源受限等特性导致现有信任模型出现计算负载、路径冗余的问题,提出了一种基于图论方法并适用于边缘计算环境的信任评估优化模型。该模型将设备间复杂庞大的信任关系抽象成有向加权图再对其进行优化处理。首先,定义了设备间的信任关系,并采用基于信息熵理论的自适应聚合方法对信任值进行聚合计算,修正了信任值之间的差异度；其次,通过添加阈值、路径长度、时间窗等多重约束条件,事先过滤掉明显不符合要求的节点和信任边,降低信任路径形成过程中不必要的计算消耗；在此基础上,利用改进后的深度优先搜索(DFS)算法对信任路径进行冗余优化处理,避免了环路以及节点绕路问题,得到优化后的信任路径集合并计算得到最终信任值。最后,将所提方法与PSM算法以及RFSN算法进行了对比实验,结果表明本文算法能够有效减轻边缘设备的资源开销,并提升了信任模型的有效性。

Optimization Scheme of Trust Model Based on Graph Theory for Edge Computing

Aiming at the problem of computational load and path redundancy in existing trust models due to resource constraints in edge computing, a trust evaluation optimization model based on graph theory,which was suitable for edge computing environment,was proposed. The model abstracted the complex and huge trust relationship between devices into a directed weighted graph and then optimized it. Firstly, the trust relationship between devices was defined, and the trust values were aggregated by using the adaptive aggregation method based on information entropy theory, which could correct the difference between trust values. Secondly, by adding multiple constraints such as threshold, path length, time window and so on, the obviously unqualified nodes and trust edges were filtered out in advance to reduce the unnecessary computational consumption in the process of trust path formation.Then, the improved depth first search (DFS) algorithm was utilized to optimize the trust path redundantly, which could avoid the problem of loops and node detour, obtain the optimized set of trust paths and calculate the final trust value.Finally, the proposed method was compared with PSM algorithm and RFSN algorithm. The results showed that the proposed algorithm could effectively reduce the resource overhead of edge devices and improve the effectiveness of trust model.