▎ 摘　　要

A comparative study on the repeater insertion designs in horizontal and vertical graphene nanoribbon (GNR) interconnects is presented. The optimal numbers and sizes of repeaters are calculated by using the particle swarm optimization algorithm, and then the results are employed to train back-propagation neural networks. By virtue of the trained neural networks, the optimal repeater designs can be quickly conducted for the GNR interconnects. The proposed optimal repeater design approach is more flexible and more adaptive than conventional analytical methods. By utilizing the proposed approach, it is found that the vertical GNR interconnects need less repeaters yet result in better performance than their horizontal counterparts.