Optimization has been the hallmark of the professional work of Jorge Moré. And the numerous awards he has won testify to the excellence of that work. For example, in 2003 he (with his colleagues) received the Beale-Orchard-Hayes Prize for Excellence in Computational Optimization. The award, given only every three years, honored his work on the Network-Enabled Optimization System (NEOS). He was named an Argonne Distinguished Scholar in 2006, and he was one of the first SIAM fellows named in 2009 for his contributions to his field
Moré received his Ph.D. in applied mathematics from the University of Maryland. He then joined the faculty at Cornell University, before coming to Argonne in 1975. An advocate of high-quality numerical software, he was on the Board of Trustees for the Wilkinson Prize for Numerical Software from its inception in 1991 until 2015. He is also coauthor of the book Optimization Software Guide, published in 1993.
In the following Q&A, Moré discusses his motivation in joining Argonne, his professional accomplishments, and the skills he considers most important for a researcher. He also gives us a look at his busy post-professional life.
What led you to join the MCS Division at Argonne, and how did it fit into your overall career goals?
I joined Argonne in 1975 because they were doing interesting work in software development. They had released EISPACK some years before I joined, and had started working on LINPACK. I was interested in developing an optimization package but did not have the support that I needed at Cornell University. The move to Argonne was risky since I had not done much formal work in software development, but Argonne had an excellent environment for software development. Keep in mind that this was in 1975, a time where the standards for high-quality numerical software were being developed.
I was also interested in continuing the theoretical analysis of algorithms that I had been doing at Cornell; I felt that the best algorithms had a solid theoretical foundation. Argonne strongly supported this point of view.
In an ideal world, I would have continued to teach, but Argonne did not have a formal teaching mechanism. I later realized that it was not possible to do all three things — teaching, software development, and theoretical analysis. Teaching requires a great deal of time, with little time for software development. Graduate students can develop software, but they are interested in graduating; software development is a means to an end. As a result the software is not polished and usually difficult to use.
What is your proudest professional accomplishment?
Trust region methods. I had been interested in a new class of methods introduced by Mike Powell that replaced the standard 1-dimensional search by a multi-dimensional search on a model. These methods had excellent performance and a solid theoretical foundation. I started working on these methods (at that time they were not called trust region methods) at Cornell University and continued this work at Argonne. In 1978 I wrote a paper “The Levenberg-Marquardt algorithm: Implementation and theory” that introduced the idea of trust region methods to least squares problems. This paper became my most highly cited paper with more than 4,200 citations.
I continued working in trust region methods in collaboration with Danny Sorensen. We extended the approach to unconstrained optimization in a 1983 paper “Computing a trust region step” that showed how to compute the step by finding the global minimizer of a quadratic subject to a ball constraint. This was an unexpected result because global minimizers are notoriously difficult to find for nonlinear problems.
Another early component of this work was the paper “Recent developments in algorithms and software for trust region methods” that I wrote for an invited lecture at the 1982 International Symposium in Mathematical Programming in Berlin. This paper has been widely credited as being influential in the development of trust region methods.
I have mentioned these papers because they are my most highly cited papers on trust region methods but I have other papers, on other topics, that have been highly cited; six of them have been cited more than 1,000 times. I am proud of these papers, often done in collaboration with others.
What skills have you developed that you believe are essential?
The ability to write manuscripts and proposals quickly and clearly is extremely important. You have to convey your ideas to researchers in the scientific community and program managers that will read your results and proposals. Much depends on the ideas presented and on the way that these ideas are presented. An attractively presented proposal is likely to be funded and thus allow you to pursue your ideas. Competition is strong so only the best ideas and proposals survive.
You also have to make presentations at conference and review panels to other researchers. The ability to make attractive and interesting presentations means that your ideas will have an audience and survive. Unfortunately, there are no courses in the typical education that teaches you these skills; you have to learn them as you progress in your professional life.
You also need to have the ability to identify and collaborate with other (the smarter, the better) researchers. I have been lucky in this respect; my collaborations with others have been highly productive and a lot of fun.
What have you been doing since retirement?
I had been thinking about retirement for a few years, so I was more or less ready when I finally decided to retire. I wanted to continue my scuba diving, but I also wanted to explore other activities that would help me stay active. Let me tell you about some of these.
Scuba diving: I like to dive from a liveaboard yacht, sleek vessels where divers live. You dive every day, usually for a week or ten days, five or four times a day. On a typical day you wake up, have a small breakfast, and then you dive. You then have a full breakfast, and dive again. Lunch and another dive is next. A snack and another dive. Dinner is next. If you are a hard core diver, you do a night dive after dinner. There is time to rest and relax between the dives, but after a day of diving, you sleep very well.
One of the advantages of scuba diving is that you get to visit interesting places. The diving sites tend to be near the equator where the ocean waters are warm. In the last five years I have been to Truk Lagoon (famous for aficionados of WW II), Palau and Yap (part of Micronesia), Fiji (an archipelago), the Sea of Cortez, the Florida Keys, Cocos Island (the island in “Jurassic Park”, also known as the Isla Nublada), the Philippines, Gran Canaria, Grand Caymans, and Thailand (Andaman Sea). My last liveaboard was in the Red Sea.
Cycling: I had always been interested in cycling, but I did not even have a bike. So, I joined a club, bought a bike, and started to learn how to bike, road biking to be precise. I like to go on longish rides, at least one hour, about 25 miles. These are not considered long rides in my club, but I tend to do these shorter rides because I ride solo. I have gone on longer rides; my longest ride is a century, 100 miles.
Yoga: I always thought that yoga was for women, that it would be too easy for me. My first yoga class showed me that I was wrong, very wrong. I was so sore after that class that I knew that this was good for me. I would recommend yoga to anybody and everybody because it is the ultimate stretching routine. I go to yoga classes twice a week. I used to be the only male in the class, but now you may find that 10% of the class is male.
Rowing: Indoor rowing is the ultimate workout, or so I am told. I keep track of distances and times rowed, and compare these with other rowers around the world. Excellent motivation and a lot of fun. I started rowing in 2011 and reached the 1 million meter rowed milestone in 2017; I am now working on the 2 million milestone.
Walking: Last year I decided to walk the Camino Francés, a 500 mile pilgrimage route from France to Santiago de Compostela in the northwest corner of Spain. I was not sure that I could do this walk in one go, so I decided to split the route into two parts. Last year I went from St Jean Pied du Port in France to Burgos, in the middle of Spain. About 200 miles in 11 days. This year I went from Burgos to Santiago, about 5,000 kilometers (300 miles) in 19 days. Each day you get up, start walking, stop for breakfast, keep on walking, have another breakfast, keep on walking, have lunch, keep on walking, have a snack, and if all has gone well, you arrive at your lodgings for the evening. I try to start walking by 6:30 in the morning and walk until 2 in the afternoon, when the sun gets really hot. I like to sleep in Albergues (a Spanish hostel) where they will fix a communal meal with other pilgrims. These meals are one of the highlights of the day since you drink good wine and exchange life stories, a profound and moving experience.