Practical FPT Implementations and Applications

Abstract

When combined with high performance computational architectures, methods born of FPT can be used as a practical basis for launching systematic attacks on large-scale combinatorial problems of significance. Efficient sequential techniques for kernelization and highly parallel algorithms for branching will be discussed. The importance of maintaining a balanced decomposition of the search space turns out to be critical to achieving scalability. Applications abound, perhaps most notably in high-throughput computational biology. A toolchain will be described that transforms immense quantities of mRNA microarray data into instances of the clique problem, which are then solved via vertex cover to derive sets of putatively co-regulated genes. This makes it possible to narrow the search for cis and trans regulatory structures on scales that were previously unthinkable.