Droplet microfluidics is likely to play a central role in the development of lab-on-a-chip technologies and as a result, significant research is directed toward this field. Understanding the spatiotemporal dynamics of discrete droplets inside microfluidic devices and the design of microfluidic devices for specific tasks are some of the dominant research topics. These works have since resulted in the development of microfluidic devices with functionalities, such as sorting, merging, synchronization, storing etc. However, the anticipated application of microfluidic devices to more complex problems will require more integrated devices that can incorporate the above functionalities on a single chip. In the current work, we present a genetic algorithm optimization-based design tool for discovering very large-scale integration of discrete microfluidic networks for a given objective function. The application of the algorithm is demonstrated through a combinatorial sequencing problem, where the objective is to achieve three different droplet combinatorial sequences for three different droplet types. Multiple fascinating, but nonobvious designs were discovered for this application. It is difficult to imagine such devices being designed using trial and error experimental procedure, which has been the main route for obtaining microfluidic device designs. With advances in technologies for fabrication of microfluidic devices, the current tool can be a significant step toward drastically cutting down on the laborious trial-and-error design process and help in developing droplet microfluidics-based lab-on-a-chip platforms cheaper and faster. © 2015 Elsevier Ltd.