Learn to Solve Constrained Routing Problems in V2V Social Vehicle Networks
IEEE Open Journal of the Communications Society · 2026 · 7 · 4660-4680
DOI: 10.1109/ojcoms.2026.3690522 ↗Kurz gefasst
Fachartikel in der Zeitschrift IEEE Open Journal of the Communications Society, 2026. Verfasst von Wei He und 4 weiteren Autor:innen. Die Arbeit nennt 78 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: National Natural Science Foundation of China.
Abstract (Originalsprache)
In vehicle social networks (VSN), multi-hop vehicle-to-vehicle (V2V) routing must minimize end-to-end delay while satisfying a social trust budget. This trust-constrained routing can be formulated as an NP-hard resource-constrained shortest path problem (RCSPP), while the per-link delay is analytically intractable because the global co-channel interference is unknown. To this end, this paper proposes a proactive infeasibility pruning (PIP) mechanism within a constrained Markov decision process (CMDP) framework. PIP constructs a budget-consistent action mask from a shortest-path-to-go prior, providing a conditional guarantee that any policy respecting the mask produces trust-feasible paths. The empirical conditioning rate of this guarantee matches the feasibility ceiling of an exact label-setting (LS) oracle to within 0.8 percentage points across the full trust-budget range. A graph neural network delay surrogate learns the interference-inclusive delay from local observations. System-level evaluation on the Wuhan Optics Valley, China, traffic demonstrates that the proposed method approaches the routing feasibility upper bound and delivers $5.3\times $ more packets than the deep reinforcement learning baselines using only local one-hop information.
Förderung
- National Natural Science Foundation of China
Zitieren
APA
He, W., Ye, J., Wang, L., Ge, X., & Han, T. (2026). Learn to Solve Constrained Routing Problems in V2V Social Vehicle Networks. IEEE Open Journal of the Communications Society, 7, 4660-4680. https://doi.org/10.1109/ojcoms.2026.3690522
Vancouver
He W, Ye J, Wang L, Ge X, Han T. Learn to Solve Constrained Routing Problems in V2V Social Vehicle Networks. IEEE Open Journal of the Communications Society. 2026;7:4660-4680. doi:10.1109/ojcoms.2026.3690522
BibTeX
@article{He2026,
author = {He, Wei and Ye, Junliang and Wang, Lijun and Ge, Xiaohu and Han, Tao},
title = {Learn to Solve Constrained Routing Problems in V2V Social Vehicle Networks},
journal = {IEEE Open Journal of the Communications Society},
year = {2026},
volume = {7},
pages = {4660--4680},
doi = {10.1109/ojcoms.2026.3690522}
}
EUVIMED