Vladimir Rokhlin is the inventor of the Fast Multipole Method for rapidly solving Poisson and Helmholtz partial differential equations. He developed many other widely used fast numerical schemes, including the nonuniformly spaced Fast Fourier Transform and randomized methods for decomposing matrices.

Abstract:
Dr. Rokhlin will discuss several new schemes for the solution of large-scale systems of linear algebraic equations, and for related tasks. The schemes are almost (but not quite) iterative, and are meant as a competitor for Conjugate Gradients and related algo-rithms. The approach will be illustrated with several numerical examples.

He received his MS from the University of Vilnius in Lithuania in 1973, and his PhD from Rice University in Houston, TX in 1983. Between 1976 and 1985, he was
an applied mathematician at the Exxon Production Research Company in Houston,
TX; since 1985, he has been working at Yale University in New Haven, CT.