Peter talked about several problems: the search for better materials for photovoltaic cells, finding the best way to place turbines on a wind farm, improving biotechnology and nanotechnology, and improving the energy efficiency of buildings. He has framed these as optimization problems but finds that standard methods aren’t always a good fit. Good search strategies are critical since it can take two days on multiple processors to calculate one data point, and there can be a million points of interest.
I learned about some optimization techniques that I might try to apply to my own efforts in finance. I was also impressed with a graphic format that can summarize high-dimensional data.