Dolf Talman is professor of Game Theory and Equilibrium Programming at Tilburg University (TiSEM) and CentER. His expertise is the development, to prove existence and the computation of equilibria in economic and game theoretic models.
During my studies Econometrics at Groningen University I was a research assistant during my last year. From that moment on I wanted to become a scientific researcher. After having finished my Master’s I became a PhD student in Operations Research in August 1976 at the Vrije Universiteit (VU) in Amsterdam.
In those days a PhD student also had to lecture every semester. For the first semester of my second year, Gerard van der Laan (a PhD student in Mathematical Economics, who had started one year before me) and I were asked by our supervisors to develop a new doctoral course. This course was focused on the intersection (not the union) of operations research (from my side) and mathematical economics (from his side). We decided to teach half a course on game theory and half a course on the computation of equilibria. This choice would determine the rest of our (academic) life!
During the course, for fifth year doctoral students in Econometrics, we thought we could improve the algorithm in the book on the computation of economic equilibria we were using. With trial and error we developed a new algorithm. During the first trials the method did not converge properly, but we were sure we would be able to make it work. Meanwhile, our supervisors became worried, because we were neglecting the research for our PhD, which was already successful. After the course they gave us two more months to succeed. If not successful, we had to continue with our original research. Luckily, we succeeded within these two months to develop what we had in mind. We wrote a paper that later got a lot of attention and was published already within one year in Mathematical Programming, a highly regarded journal. Our supervisors were convinced and we still had one year to continue our joint research to generalize our new method. We also found a supervisor from Northwestern University, because our supervisors were not very familiar with the topic.
During that summer we presented our work at several places, even at a conference where we were not registered as speakers, and at the Econometric Society European Meeting in Geneva. Our presentations were in the last session of the last day of the meeting and the two other talks were in French, so there was almost nobody present. Except for the speakers and chairman and a few other people. Herbert Scarf however was present, the author of the book we had been using, and also Pieter Ruys, from Tilburg University. Both became later my bosses. Within one year we completed seven other papers and split these eight papers to write two PhD theses. This division, which was not completely arbitrary, was after all a fair division, because both theses contained exactly the same number of pages. At that time writing a thesis took one year, because the secretaries had to type it on an IBM electronic typing machine and making corrections was almost impossible. We defended our PhD theses in a single joint academic session on February 29, 1980.
Our cooperation continued after our PhD defenses, although I first went one year to Yale University in New Haven (see my previous column) as visiting lecturer under Herbert Scarf and then became assistant professor under Pieter Ruys at Tilburg University. Over the years we wrote dozens and dozens of papers together both on the computation and existence of equilibria and, later on, on cooperative game theory. We supervised PhD students, also several jointly, visited many universities, from Stanford to Tokyo, and attended lots of conferences all over the world to present our research.
Text by: Dolf Talman