Julian Hall
Michael Forward has written the following article as part of our series of Academic Interviews; featuring Julian Hall!
Dr. Julian Hall is a member of the School of Mathematics’ Optimization and Operational Research Group, and is currently working on the open-source linear optimization software known as HiGHS. Much of his professional academic work has focused on developing computational techniques to enable the classical revised simplex method to solve linear programming problems more efficiently. Following the work of a previous PhD student, he now manages the development of HiGHS which aims to generate impact within industry and the academic world. Julian has done this in the past with his previous linear programming solver, which is currently used to calculate the “optimal raw material blend for about fifty percent of the world’s manufactured animal food.” With such an impressive statistic, I was keen to trace his path from undergraduate study to his work today and find out what keeps him occupied.
Julian attended the King’s School in Macclesfield from 1974 to 1983, and sat his Oxford entrance exam after doing A-levels. In the months between leaving school and going up to Oxford, he worked for the pharmaceutical research company Imperial Chemical Industries (ICI), now known as AstraZeneca. It was here that Julian first encountered optimization. Tasked with solving an optimization problem, he was told that he didn’t need to come up with an algorithm, and was given a rather intriguing book on the subject. He was hooked. Although studying optimization with its author, Roger Fletcher, was never a “dream”, he would eventually become Julian’s PhD supervisor. Julian’s time at ICI proved useful in preparing him for undergraduate study, as he left with a much clearer idea of where his interests and abilities lay, allowing him to focus on those areas as he moved through the university. He’d always anticipated going into industry after studying at Oxford but was, by now, fully enthused by Numerical Analysis and Optimization. He decided that he wanted to learn more, and began an MSc programme in this area at Dundee, before switching to a PhD under Roger.
For the next 25 to 30 years Julian’s main research focus was on solving linear programming problems using the simplex method. To massively simplify what the simplex method involves, it is essentially a process requiring the solution of a sequence of systems of related linear equations, and has been in use for around 75 years. During this time Julian wrote his own linear programming solver and, for many years, it has been used in animal feed formulation software produced by Format Solutions and in oil reservoir management software from Weatherford. With barely a hint of lamentation, he admits that his solver was bested by a PhD student of his, and it was this solver that eventually grew into the HiGHS software that he is currently developing with current PhD students. In fact, HiGHS is planned to replace the solver used in Format Solutions’ software, to determine the optimal composition of tens of billions of dollars’ worth of animal food each year.
Due to the nature of his work, Julian gets to speak to, and solve problems for, lots of different people, many of whom lie outside academia. For Julian, this is perhaps the most fulfilling thing that his job entails. He stresses the importance of being able to create impact in the wider world, so learning about what kind of situations non-academics are modelling, and feeding that back into his own work, is very rewarding. More recently, Julian has received funding from Google and Huawei, so exciting things are on the horizon for sure.
Writing 15 papers in 32 years might sound insignificant, but four have won their journal’s best paper prize (2005; 2013; 2015; 2018), and Julian’s focus on coding and consultancy has led to the creation of valuable impact. Although he hopes that his example is refreshing, and perhaps inspiring to some, he regrets that prowess in research is seen by many as the only objective of an academic career. For Julian, applied mathematics is the way forward — not just for the personal joy of studying it but for its ability to create large-scale, material impact on the world around us.