Exploiting sparsity in Jacobian computation using graph coloring

MSc MTech Krishan Gahalaut

Jan. 26, 2010, 2:30 p.m. P 215

The computation of the derivative information is fundamental to many numerical algorithms. For example, in general, the minimization of a function requires the gradient and/or Hessian of the function, or, the solution of a system of nonlinear equations requires the Jacobian matrix. Based on the sparsity exploiting techniques, this work deals with efficient computation of large Jacobian. We will discuss how to take advantage of the sparsity using graph coloring.