About the Author

Obituary: Thomas F. Coleman

By Alex PothenStephen Vavasis, and Henry Wolkowicz

Thomas F. Coleman, 1950-2021. Photo courtesy of the authors.
Thomas F. Coleman—a leader in optimization and scientific computing, professor at the University of Waterloo, and a SIAM Fellow—passed away after a battle with cancer on April 20, 2021 in Waterloo, Canada.

Tom was born on November 21, 1950 in Montreal, Canada and attended high school in Toronto. Thanks to the influence of a high school math teacher, he matriculated at the University of Waterloo via a co-op program and spent alternating semesters at school and in industry. After completing his Ph.D. in 1979 under the supervision of Andy Conn in Waterloo’s Department of Combinatorics and Optimization, Tom spent two years as a postdoctoral researcher at Argonne National Laboratory. He then held a faculty position in the Department of Computer Science at Cornell University from 1981 to 2005 and served as director of the Cornell Theory Center (now the Cornell Center for Advanced Computing) from 1998 to 2005. Tom returned to Waterloo in 2005 as dean of the Faculty of Mathematics and a professor of combinatorics and optimization. He served as dean until 2010.

Tom’s early papers—several of which he wrote jointly with Conn—addressed the convergence of nonlinear programming with exact penalty functions. While at Argonne, Tom began to investigate combinatorial algorithms in optimization and scientific computing. He coauthored a series of papers with Jorge Moré on the use of graph coloring for efficient sparse Jacobian and Hessian estimation. Tom realized that ideas from combinatorial optimization could significantly improve existing techniques for large-scale computation and pursued this line of research throughout his career. Sparse matrix computation has long been a source of interesting problems that link combinatorial and scientific computing, and Alex Pothen—Tom’s first Ph.D. student—tackled the problem of finding a sparse null space basis in his dissertation. Tom guided several students in this research space; one might consider him the father of the area now called “combinatorial scientific computing” (CSC). In 2018, researchers in CSC joined with other research sub-communities to form the SIAM Activity Group on Applied and Computational Discrete Algorithms (SIAG/ACDA), and the first SIAM Conference on ACDA will take place virtually this July. 

In the 1990s, Tom developed the interior trust-region method jointly with Yuying Li. This work became part of the optimization toolbox in MATLAB and comprises Tom’s most highly cited paper. Automatic differentiation is another computational issue that drew Tom’s attention during this period. He foresaw the significance of automatic differentiation for optimization and promoted the subject with publications, software tools, a book, and conferences. 

Around this same time, Tom developed an interest in optimization in the financial sector and wrote an influential sequence of papers on financial derivatives and portfolio optimization. He remained active in this area until the end of his career. In addition to his publications, Tom organized more than a dozen conferences, workshops, and lecture series on the topic of financial optimization. He also founded a consulting and software company with his son Conrad called Cayuga Research, which provided an outlet for more direct commercialization of his ideas.

Equal to his scholarship contributions were Tom’s leadership roles at Cornell and Waterloo. For example, he sharpened the Cornell Theory Center’s focus on research in scientific and financial computing while serving as its director. Tom also opened the center’s Wall Street branch in 1999, foreseeing the demand of Wall Street firms for the latest academic research.

Tom was appointed dean of the Faculty of Mathematics at Waterloo in 2005 and oversaw five academic departments as well as numerous programs and institutes. Among his first acts as dean was opening a Waterloo office on Wall Street. A delegation that included Tom and other Waterloo leaders celebrated this event by ringing the stock exchange’s starting bell on November 23, 2005 — an honor usually reserved for visiting heads of state and CEOs. Tom’s leadership inspired a number of achievements at Waterloo: the math business programs grew, the school created a graduate program in computational mathematics, the faculty obtained funding for a new math building (the third on campus), and the Centre for Education in Mathematics and Computing (CEMC) obtained a large grant from the Gates Foundation. The CEMC has become the largest mathematics outreach organization in Canada and is known worldwide for its educational materials and grade- and high-school math contests.

After stepping down as dean in 2010, Tom remained active in research, teaching, and administration until he became ill in the fall of 2020. He served a five-year term as director of the Waterloo Research Institute in Insurance, Securities, and Quantitative Finance and held the Ophelia Lazaridis University Research Chair at Waterloo.

Tom was known for his careful preparation and thoughtful decision-making throughout his career, which allowed him to foresee many major trends in mathematics and computing. He mentored 21 Ph.D. students, 18 postdoctoral fellows, and 18 master’s students; these students now hold leadership positions in academia, the U.S. Department of Energy’s national laboratories, and industry — especially in the financial sector. Tom’s students remember him as an exceptional mentor who was available to discuss research in a supportive manner and who exuded an optimistic and practical outlook. He empowered students to think independently and encouraged them to take risks in research. By working hard, playing hard, and developing close personal relationships with his students—all seasoned with a fine sense of humor—Tom served as a role model for living a full and balanced life. Upon hearing the news of his illness, many of his former students contacted Tom to let him know about the profound influence he had in their lives. 

Tom was an active athlete throughout his life and enjoyed cycling and canoeing with his family. He is survived by his wife Yuying Li (a colleague in Waterloo’s Computer Science Department); his children Conrad, Dan, and Lena; and his former wife, four siblings, and several nieces and nephews.

More information about Tom’s life and career is available on his University of Waterloo legacy page. A minisymposium at the 2021 SIAM Annual Meeting will address Tom’s contributions to applied mathematics. 

Alex Pothen is a professor of computer science at Purdue University and was privileged to be Tom Coleman’s first Ph.D. student at Cornell University. Stephen Vavasis is a professor in the Department of Combinatorics and Optimization at the University of Waterloo and was a colleague and friend of Tom at both Cornell and Waterloo since 1989. Henry Wolkowicz is a professor in the Department of Combinatorics and Optimization at the University of Waterloo. He was fortunate to become acquainted with Tom in 1979 just as he was finishing his Ph.D. They remained friends and colleagues and were members of the same department and faculty since 2005.