Students Jonah Aldridge and Joe Carstairs have worked together to produce this article as part of our series of Academic Interviews; featuring Lars Schewe!
Open Grid Europe (OGE)
Open Grid Europe (OGE) is Germany’s dominant gas transport network. They are responsible for keeping gas infrastructure, such as pipelines, in good working order, and renting their use to distributors; much like SGN does here in Scotland. However, when millions of businesses and households depend on the delivery of gas through 12,000km of pipes, roughly equivalent to the length of Germany’s entire motorway network, predicting how much gas will be needed on any one day, and feeding it into the system accordingly, is no trivial task. This is where Dr Lars Schewe comes in. In 2016, together with a team of researchers at various German universities, he tackled the problem head-on.
OGE’s optimization problem
Dr Schewe works in optimization, one of mathematics’ major fields and among its most important in our increasingly data-driven world. Optimization is the art of modelling a complex system and finding a way to maximise the quantities you most desire, while minimising the quantities you would rather do without. This is why OGE’s problem was an optimization one: they needed to maximise how much gas they could supply across the network while minimising the need to transport it across long distances at great expense.
To give a sense of scale to the problem, Dr Schewe and his team were modelling a vast, complicated physical system stretching right across Germany. They had to account for a huge variety of variables, from how the weather would affect household demand for heating, to the rate at which gas pressure declines as it travels through a pipe. Worse still, many of the physical laws governing gas are non-linear. When it came to non-linear relationships, they had a choice between approximating those relationships with linear ones, or simulating them more accurately at the expense of using exponentially greater computing power.
Real-life optimization problems, Dr Schewe explains, are rarely about finding a solution. Academics have found theoretical optimization problems which are unsolvable, but such problems are not likely to occur in the physical world. Instead, the work of optimizers like Dr Schewe is not in finding “the” solution, but rather finding the most efficient one.
Studying at University
Dr Schewe feels quite at home in optimization today, but it wasn’t always clear that he would work in such an applied area of mathematics. Indeed, it wasn’t always clear that he would be a mathematician at all; at undergraduate level, he studied a joint degree in Mathematics and Sociology. However, it was at PhD level that he started to turn towards optimization. His topic of research, in discrete geometry, might have been abstract; but he increasingly saw a divide amongst mathematicians, between “grand theorists” on the one side and “problem-solvers” on the other. He decided that he belonged in the latter group, and when the opportunity in Erlangen-Nürnberg arose, he jumped at the chance.
The University of Edinburgh and his Motivations
Since then, he’s made the move to the University of Edinburgh; attracted by the world-class optimization group here. Having moved from working with OGE to the National Grid, Dr Schewe has shown that he’s fascinated by one common truth: working in optimization puts him at the cross-section of the physical sciences and political economy. He gets to see the way that laws, regulations and market forces interact in one of the most complex systems of all: the one that delivers goods and services to us, the consumers. Not all of that machinery is very pretty, but by engaging in dialogue with both physical scientists and economists, he hopes to keep on with what mathematics does best: solving problems.