The INFORMS awards the George B. Dantzig Dissertation Award for outstanding dissertations on the interaction between academia and industry. Awardees have demonstrated innovative research relevant to practice in operations research or management science. Daniel Rehfeldt made it to the finals of 2022 for his thesis on “Faster algorithms for Steiner tree and related problems: From theory to practice.” The other finalists were Anish Agarwal (MIT), Ruihao Zhu (MIT), and Su Jia (Carnegie Mellon University), who was awarded for his thesis on “Learning and Earning Under Noise and Uncertainty”. The award was given at the INFORMS Annual Meeting held in Indianapolis, October 16th to 19th.
The Steiner tree problem in graphs (SPG) is one of the most studied problems in combinatorial optimization. Many applications can be modeled as SPG or closely related problems. In his work, Daniel introduced many new algorithmic components for solving SPG, such as reduction techniques, cutting planes, graph transformations, and heuristics – for both SPG and 14 related problems. Many of these methods and techniques are shown to be more effective than previous results from the literature. The Steiner tree solver SCIP-Jack (with source code freely available for academic use https://scipjack.zib.de/ ) developed in this work is faster than all other solvers from the literature (including problem-specific ones) in each of the 15 problem classes, often by orders of magnitude. SCIP-Jack is being used in several industrial projects, including the design of highspeed-internet networks in Germany.
Congratulations to Daniel and the other finalists.