Summary
Scientific simulations now routinely generate data streams vastly in excess of the ability of humans to interpret. Recent work in data analysis and visualisation has therefore applied topological analysis to the task of feature detection and interpretation. However, most of the algorithms are essentially serial, although some work has been performed on parallelisation. The goal of this PhD would be to identify common patterns across multiple topological computations, and identify how to parallelise not only one single approach, but as many as possible, by exploiting fundamental properties both of the algorithms and of the underlying mathematics that they encode.
