Partial Differential Function Integrated Generative Learning for Reducing Intermittent Message Disseminations in Vehicular Ad-Hoc Networks Connectivity

Vehicular Ad-Hoc Networks (VANETs) are described by their volatile connections, rapid network dynamics, and frequent neighbor changes. Due to these characteristics, message dissemination is unreliable, and periodic neighbor discovery increases the routing overhead. To provide a reliable solution for this problem, this article proposes a Partial Differential Generative Learning (PD-GL) for a connectivity assessment model. The differential connections between the vehicles are derived using the partial function, followed by the connection probability assessment using generative learning. In this process, the change in network dynamics initiates the partial differential function to verify communication links between the vehicles. This probability assessment is repeated until the maximum driving distance of the vehicle is supported by communication. The change in network dynamics using the GL identifies low probability vehicles to pre-exclude from the communication, reducing the overhead. The proposed PD-GL model improves the message delivery by 13.61% and connection probability by 13.84%. This model reduces the dissemination overhead by 12.68%, connection loss by 13.29%, and dissemination delay by 12.61%, for the maximum acceleration.