An Artificial Neural Network Surrogate Model for Repeater Optimization in the Presence of Parametric Uncertainty for Hybrid Copper-Graphene Interconnect Networks

In this paper, an artificial neural network model is developed to predict the statistics of the optimal number and size of repeaters required to minimize the power delay product (PDP) of on-chip hybrid copper-graphene interconnect networks when subject to parametric uncertainty. The proposed ANN model is a composite of two smaller ANN models. One ANN model is used to emulate the per-unit-length parameters of the interconnects as functions of the geometrical, physical, and material parameters of the network. A second ANN model takes as inputs the outputs of the first ANN model and predicts the corresponding optimal number and size of the repeaters required in the network. Overall, the composite ANN model enables the use of analytic expressions instead of expensive and repeated full-wave electromagnetic (EM) simulations to solve the repeater optimization problem. This composite ANN model is used in a Monte Carlo framework for efficient statistical analysis.