Retracted: Path Selection for Packet Transmission in Mobile Devices Communications by Deep Learning Technique

The proposed neural network's objective is to reduce network time delay within the context of this study's optimization of the packet pathways. The routing algorithm's primary concern, which can be resolved with real-time path computations, is said to be finding the shortest route. It is possible to achieve excellent performance by taking advantage of the gaps in earlier research, which are reflected in the system's insufficient training and the erroneous predictions as a result of not accounting for the feedback from the hidden layers. In order to enhance the current approaches employing weights obtained from packet ID and to simultaneously alter neural network iteration, this study attempts to propose an effective technique. In this paper, the creation of an effective neural network with the right output is covered in detail, along with the network's underlying concepts. The study's conclusions showed how efficient it is to use computational systems to showcase computer simulation. The study's conclusions showed that it is highly effective to use the computational system's energy to show off numerical simulations. It is also demonstrated that the system performed well when trained to obtain a 3 percent time delay with 5 nodes in mobile device communications. In order to determine whether to permit each D2D pair to transmit data, a CNN is constructed. We gather data samples from a suboptimal algorithm in order to train the CNN and verify the trained CNN. Due to the limited computer capacity, our numerical findings demonstrate that the accuracy of the proposed deep learning-based transmission technique reaches roughly 85%– 95% despite its simplistic structure. Additionally, the outcomes demonstrated that the key components of the suggested model may operate in real time and are adaptable to changes in path topology.