Placement Algorithm for Flow-Based Microfluidic Biochips

Microfluidic biochips revolutionize biology by placing laboratory functionality on a very small chip. The subject of this thesis is a flow-based biochip, which consists of thin channels in which fluids flow. The basic building block of the chip is a valve, which can direct the flow by blocking or unblocking channels. The valves are combined to create components like switches, mixers, and mul-tiplexers. Biochips are currently designed manually using drawing tools like AutoCAD. The process is both tedious and error prone. Especially as biochips become larger and more complex. The biochip design process consists of many phases. One of the first things that must be decided, is the placement of the components. The designer seeks the best placement as he places each component on the chip. Components should be packed close together to reduce the chip size. The placement should also attempt to facilitate direct and non-intersecting routes between components. This is to minimize the flow transportation time between components and to avoid introducing additional switches. This thesis proposes an automated component placement tool, which assists the designer in choosing the best placement. The following chapters describe the algorithmic aspects of analysing and implementing such a tool. Components, connections, routes, and the chip itself are modelled to represent a component placement. A heuristic algorithm is proposed for optimizing the placement, and finally the algorithm is evaluated on several synthetic and real life benchmarks.