A novel use of deep learning to optimize solution space exploration for signal integrity analysis

Abstract

The enhanced complexity of electrical devices has increased the number of variables that directly or indirectly affect the output. Consequently, it has become imperative to explore the massive solution space during integrity analyses, without sacrificing accuracy and development time. In this paper, we offer a simple hybrid algorithm based on a Multi-Layer Perceptron that significantly works better than traditional methods like Least Squares, by balancing the requirements for high accuracy and less development time.