Placement-Aware Architectural Synthesis of Digital Microfluidic Biochips using ILP

Microfluidic-based biochips are replacing the conventional biochemical analyzers, and are able to integrateonchip all the necessary functions for biochemical analysis u sing microfluidics. The digital microfluidic biochips are based on the manipulation of liquids not as a continuous flow, but as discrete droplets (hence the termdigital), and thus are highly reconfigurable and scalable. We model a biochemical applica tion using an abstract model consisting of a sequencing graph. Th e digital biochip is modeled as a two-dimensional array of cel ls, where each cell can hold a droplet. In this paper we propose an integer linear programming (ILP) synthesis methodology that determines the allocation, resource binding, and sche duling of the operations in the application (architectural synthesis) at the same time with module placement (physical synthesis) . Although architectural and physical synthesis steps are ty pically performed separately, we show that significant improvement s can be gained by considering the placement during architect ural synthesis. To handle large problem sizes, we have extended t he ILP implementation with Local Branching, which is a metaheuristic for design-space exploration that uses the ILP so lver to perform the local searches. The proposed methodology has been evaluated using several real-life examples.