Early-stage microfluidic network design framework using graph sparsificiation based optimisation

Microfluidics has emerged as an effective technology for miniaturised manipulation and biochemical analysis. Recently, several physical design solutions for functional microfluidic chips have been presented. However, the commonly designed grid structure happens to be redundant with unnecessary channels, causing significant waste of expensive sample fluid. In this work, the authors propose an early stage microfluidic network design framework to sparsify the network topology and optimise network area overhead. Simulation results show that the proposed framework can generate the network to meet design constraints and reduce design overhead by almost 79.2% compared with prior works.