On Operations Research Pioneer Richard Bellman *46
As an undergraduate classics major who went on to graduate and doctoral degrees in operations research and management science, I enjoyed hearing about Katy Milkman ’04’s application of OR to 1990s New Yorker literary fiction in her senior thesis. It reminded me of a passage from the memoir of the great Princeton mathematician and OR pioneer Richard Bellman *46. Bellman is most famous for his important contributions to dynamic programming and control theory, which also are foundational to today’s reinforcement learning algorithms. In his Eye of the Hurricane, he discusses his time as a graduate student and assistant professor at Princeton, before going on to the RAND Institute, where much of modern OR was developed, recounting a string of adventures and troublemaking along the way, in the manner of Richard Feynman *42 (with whom we worked on the Manhattan Project). He later returned to work on the top-secret Project Matterhorn that became the basis of today’s Princeton Plasma Physics Laboratory. Probably all ORFE students have encountered the Bellman-Ford shortest path algorithm. In his memoir, the brilliant but mischievous mathematician recommends that operations research courses teach Mark Twain’s hilarious essay “Fenimore Cooper’s Literary Offenses,” a surprising recommendation. Although he does not specifically say why, the essay was both a parody but also instantiation of quantification applied to literature. For instance, a typical passage runs, “In one place in Deerslayer, and in the restricted space of two-thirds of a page, Cooper has scored 114 offenses against literary art out of a possible 115. It breaks the record.” The subtle point seems to be, which Bellman made frequently, that effective mathematical models have to be sufficiently complex, but not overly so (indeed he coined the concept of “the curse of dimensionality”), a lesson which is at the heart of OR.
As an undergraduate classics major who went on to graduate and doctoral degrees in operations research and management science, I enjoyed hearing about Katy Milkman ’04’s application of OR to 1990s New Yorker literary fiction in her senior thesis. It reminded me of a passage from the memoir of the great Princeton mathematician and OR pioneer Richard Bellman *46. Bellman is most famous for his important contributions to dynamic programming and control theory, which also are foundational to today’s reinforcement learning algorithms. In his Eye of the Hurricane, he discusses his time as a graduate student and assistant professor at Princeton, before going on to the RAND Institute, where much of modern OR was developed, recounting a string of adventures and troublemaking along the way, in the manner of Richard Feynman *42 (with whom we worked on the Manhattan Project). He later returned to work on the top-secret Project Matterhorn that became the basis of today’s Princeton Plasma Physics Laboratory. Probably all ORFE students have encountered the Bellman-Ford shortest path algorithm. In his memoir, the brilliant but mischievous mathematician recommends that operations research courses teach Mark Twain’s hilarious essay “Fenimore Cooper’s Literary Offenses,” a surprising recommendation. Although he does not specifically say why, the essay was both a parody but also instantiation of quantification applied to literature. For instance, a typical passage runs, “In one place in Deerslayer, and in the restricted space of two-thirds of a page, Cooper has scored 114 offenses against literary art out of a possible 115. It breaks the record.” The subtle point seems to be, which Bellman made frequently, that effective mathematical models have to be sufficiently complex, but not overly so (indeed he coined the concept of “the curse of dimensionality”), a lesson which is at the heart of OR.