University of Connecticut

Numerical Analysis

Research Activity

  • Numerical solutions of partial differential equations, especially those arising in applications in science and engineering. Emphasis is on finite element methods.
  • Scientific computing, particularly in the design of fast and superfast algorithms for structured matrices (e.g., Toeplitz, Hankel, Cauchy, Pick, semiseparable, quasiseparable and multilevel matrices) with applications to problems arising in electrical engineering (e.g., signal processing, system theory, filtering and estimation, algebraic coding theory, stability and control theory).

Particular attention is devoted to the error analysis of algorithms and numerical accuracy issues.