Professors Adrian Lewis, Eva Tardos and Michael Todd Named to the Inaugural "Class" of SIAM Fellows
Known in industrial and academic mathematical circles as SIAM, the Society for Industrial and Applied Mathematics recently established a Fellows program to recognize the work of its distinguished members. ORIE Faculty members Adrian Lewis, Michael Todd and Éva Tardos, were among those in the initial group of Fellows. According to the SIAM web site, the initial Fellows were "selected from among those SIAM members whose previous levels of recognition place them clearly among the members intended to be recognized by the program." Lewis, Tardos and Todd are among the eight Cornell faculty members chosen.
Lewis joined ORIE from Simon Fraser University in Canada as a full professor in 2004. His B.A., M.A. and Ph.D. degrees are from the University of Cambridge. Todd, the Leon C. Welch Professor of Engineering, also holds a B.A. from Cambridge. His Ph.D. is from Yale University. He joined the Cornell faculty in 1973. Tardos, who is currently the chairperson of the Department of Computer Science, was named a Jacob Gould Schurman Professor in 2007. She received a Ph.D. from Eötvös University in Hungary and joined the Cornell faculty in 1989.
All three SIAM Fellows work in areas related to the problem of optimal allocation of scarce resources, encompassed by the term optimization, or mathematical programming. Todd's work focuses on so-called interior point methods, which received prominence with the development of the Karmarkar algorithm for linear programming in 1984. Such methods can be generalized to classes of resource allocation problems in which, unlike linear programming problems, resource consumption, costs and/or profits are nonlinear (i.e. output or resource consumption is not additively proportional to input or activity level). Todd has worked on a class of such optimization problems that is particularly well-suited to such generalizations, called semidefinite programming.
Tardos has focused on efficient algorithms for network optimization problems. Early on, she developed a particularly effective approach for solving the so-called transportation problem, in which a quantities of a commodity are to be delivered, at least cost, from a set of supply points to a set of demand points. Her algorithm settled a long-standing open question about whether such an efficient (polynomial time) method could be developed. More recent work has shown the power of mathematical programming approaches in the design of approximation algorithms for a broad cross-section of problems including facility location, clustering, and scheduling problems. Most recently she has worked in a relatively new field called algorithmic game theory, and in particular on the design and analysis of algorithms for situations involving individual optimization by selfish users or players.
Lewis is interested in the mathematical analysis and computational solution of complex design optimization problems. A simple engineering example is the design of a shock-absorber: by varying the grade of lubricant, the designer can modify how stiff the system feels. Finding optimal choices for damping vibrations in such a physical, electronic, communications or information system is "nonsmooth": choices are very sensitive to slight changes in the system, one reason they seem hard to compute. Lewis facilitates this important but challenging optimization process by blending mathematical modeling and analysis with scientific computation techniques.
For additional information, click here.