Computer algorithm design and optimisation involves, in particular, constructing sequential and parallel algorithms, analysing computational and memory complexities of algorithms, as well as optimizing algorithms and deriving trade-offs. These include both centralized, on-line and distributed algorithms, especially in the field of graph theory. In the distributed setting, mobile agent computing is of particular interest. This involves graph exploration, leader election, and general searching methods with potential application to navigation algorithms for teams of robots. Other applications involve traffic routing methods (graph-based algorithms, maze routing algorithms, evolutionary algorithms, genetic algorithms, ant colony algorithms etc.), which involve control of automated vehicles, planning or determining collision avoidance manoeuvres, collision risk measures and weather-based routing. Furthermore, this field of research includes parallelization of algorithms, design and implementation of code targeting modern computer systems, including high performance computing involving both shared and distributed memory, with multi- and many-core CPUs, accelerators including GPUs, as well as hybrid systems including clusters of CPU+GPU nodes. Other research items involve making use of optimizations including (dynamic) load balancing, minimization of synchronization overheads and idle times, hiding communication latencies, using optimized memory access patterns, tiling etc. Finally, the field also includes building, optimizing and deployment of small and large-scale software frameworks.
Keywords:
algorithm, optimization, routing, high performance computing systems, multi-core CPUs, many-core CPUs, GPUs, clusters
