EUVIMEDEUVIMED entdecken
Crossref

Simulation Modeling and Evaluation of Node Reputation in P2P Networks

Volodymyr Mikhav, Yelyzaveta Meleshko, Mykola Yakymenko, Oleksandr Didyk, Oleksandr Sereda, Anton Subbota

Vìsnik Nacìonalʹnogo unìversitetu "Lʹvìvsʹka polìtehnìka". Serìâ Ìnformacìjnì sistemi ta merežì · 22.06.2026 · 19 · 483-502

DOI: 10.23939/sisn2026.19.483 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Vìsnik Nacìonalʹnogo unìversitetu "Lʹvìvsʹka polìtehnìka". Serìâ Ìnformacìjnì sistemi ta merežì, online erschienen am 17.05.2026. Verfasst von Volodymyr Mikhav und 5 weiteren Autor:innen. Die Arbeit nennt 81 Literaturangaben.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

The article proposes a simulation model of node reputation in a peer-to-peer network based on local evaluations of interaction outcomes between network participants and their aggregation into a global reputation using the EigenTrust method. Interactions between nodes are modeled as the exchange of services or resources, while the behavior of each node is defined by a cooperation honesty parameter that accounts for behavioral noise. Based on locally accumulated evaluations, a trust matrix is constructed and used to iteratively compute global reputation. The paper presents a comparative analysis of the model’s operation under two interaction graph topologies of network participants: a fully connected network and a complex network based on the Barabási–Albert model. In addition, a modification of the EigenTrust method is proposed, namely, a node inclusion mechanism in the global reputation space based on a minimum social coverage threshold. The quality of global reputation estimation was evaluated using the Pearson and Spearman correlation coefficients, MAE and RMSE errors, as well as the ranking metrics Overlap@k, Jaccard@k, and AUC. The experimental results showed that the structure of the interaction graph of network participants is a key factor influencing the quality of global reputation estimation. In the experiments, the fully connected graph yielded an almost perfect correspondence between global reputation and the true honesty of nodes. In contrast, for the graph based on the Barabási–Albert model, the baseline reputation model showed only moderate accuracy and a pronounced bias whereby high global reputation values are assigned to nodes with a large number of interactions regardless of their actual honesty. It was found that introducing a mechanism for admitting nodes into the global reputation space with a minimum social coverage threshold of 20% of the network increases the correlation between global reputation and true honesty for the interaction graph based on the Barabási–Albert model. Therefore, the obtained results indicate that for heterogeneous P2P networks, a promising avenue for improving reputation systems is the development of mechanisms for the preliminary filtering of nodes with insufficient reputation evidence

Zitieren

APA

Mikhav, V., Meleshko, Y., Yakymenko, M., Didyk, O., Sereda, O., & Subbota, A. (2026). Simulation Modeling and Evaluation of Node Reputation in P2P Networks. Vìsnik Nacìonalʹnogo unìversitetu "Lʹvìvsʹka polìtehnìka". Serìâ Ìnformacìjnì sistemi ta merežì, 19, 483-502. https://doi.org/10.23939/sisn2026.19.483

Vancouver

Mikhav V, Meleshko Y, Yakymenko M, Didyk O, Sereda O, Subbota A. Simulation Modeling and Evaluation of Node Reputation in P2P Networks. Vìsnik Nacìonalʹnogo unìversitetu "Lʹvìvsʹka polìtehnìka". Serìâ Ìnformacìjnì sistemi ta merežì. 2026;19:483-502. doi:10.23939/sisn2026.19.483

BibTeX
@article{Mikhav2026,
  author = {Mikhav, Volodymyr and Meleshko, Yelyzaveta and Yakymenko, Mykola and Didyk, Oleksandr and Sereda, Oleksandr and Subbota, Anton},
  title = {Simulation Modeling and Evaluation of Node Reputation in P2P Networks},
  journal = {Vìsnik Nacìonalʹnogo unìversitetu "Lʹvìvsʹka polìtehnìka". Serìâ Ìnformacìjnì sistemi ta merežì},
  year = {2026},
  volume = {19},
  pages = {483--502},
  doi = {10.23939/sisn2026.19.483}
}
Kostenlos starten →