ARMM: Adaptive Reliability Quantification Model of Microfluidic Designs and Its Graph-Transformer-Based Implementation

After decades of development, flow-based microfluidic biochips have become a revolutionary platform for biochemical experiments. To meet the increasingly complex experimental demands, the length and density of channels in these chips grow significantly, which brings about higher defect probabilities. Till now, several methods have been proposed to improve the yield of these increasingly complex chips. However, the effectiveness of these methods cannot be properly evaluated, since there has been no method that systematically analyzes the reliability of a microfluidic design. In this paper, we propose the first mathematical models to quantify the reliability of a microfluidic design by calculating the probability of blockage and leakage defects happening to the design. Besides, we propose a graph-transformer-based method to speed up the calculation, so that designers can have a fast and accurate evaluation of the reliability of a microfluidic design at any scale.